# Research Papers — 2026-09 · bundle page 39 of 266 > Full-text markdown of papers #1901–1950 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/39 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/38.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/40.md) ### Epistemic Policy Divergence in Multi-Turn LLM Contamination: A Protocol-Gradient Investigation Authors: Fahrell Giovanny, Geby Bayuningtyas, Sahrul Mukharom, Hafiz Budi Firmansyah Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35308v1 Large language models process conversation history as unverified context: false premises injected into prior turns can be adopted as fact, a failure mode we term session-level contamination. We introduce five contamination protocols arranged along a source-authority gradient, isolating distinct failure mechanisms while holding the false premise constant, and evaluate GPT-5.4 Mini, Gemini-3.1 Flash-Lite, and GLM-4.5-Air across ten knowledge domains at temperature zero (22,500 turns), using a dual-track automated judge validated against a human gold standard (Cohen's \k{appa} = 0.901). GPT-5.4 Mini showed zero adoptions across all 500 sessions, a content-independent policy at the session level; token-level probing shows the underlying margin, while large, is finite. Gemini-3.1 Flash-Lite followed a steep authority gradient: 0.1% adoption for self-attributed falsehoods, 23.5% for user-cited sources, 68.2% for system-injected authority, and 94.0% under instruction override. GLM-4.5-Air showed a shallower gradient (15.8% vs 84.2%), a 68-percentage-point dissociation confirming that authority deference and instruction compliance are distinct mechanisms within one architecture. Recovery also diverged: GLM recovered in 94.5% of affected sessions, whereas 26.1% of affected Gemini sessions never did, rising to 40.0% under instruction override. Conversation history is an untrusted attack surface requiring provenance-aware system design; the complete framework is released as an open-source benchmark. --- ### PIVOT: Pivot-Aware On Policy Self Distillation for Multi-Turn VLM Agents Authors: Jiazhou Zhou, Hu Zhou, Yucheng Chen, Jinyuan Qu, Ying-Cong Chen, Lei Zhang Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35303v1 Reinforcement learning with verifiable rewards (RLVR) via Group-Relative Policy Optimization (GRPO) is widely used for multi-turn VLM agent training, yet it suffers from zero-gradient silence on uniform failures and coarse episode-level credit assignment. While On-Policy Distillation (OPD) and On-Policy Self-Distillation (OPSD) mitigate sparse rewards using hindsight information, their underlying mechanisms remain poorly understood. Through controlled counterfactual rollback probes across five multi-turn VLM agent benchmarks, we reveal that performance gains in OPSD/OPD are largely driven by physical state rollback at the pivot step, defined as the first unrecoverable action without remaining step budget. However, physical state rollbacks are computationally prohibitive and infeasible in real-world environments. To bridge this gap, we present Pivot-Aware Internalized Visual On-Policy Training (PIVOT), an RL framework that internalizes pivot localization and state restoration directly into token-level parameter updates, eliminating environment rollbacks during RL training and additional skill hints at test time. PIVOT unifies three functional roles within a single architecture: a failure Analyzer non-invasively localizes the pivot step and diagnoses failure modes from visual trajectory collages and action logs; a detached Teacher re-scores failed tokens under this privileged diagnostic context; and a Student optimizes joint GRPO and confidence-gated OPD objectives. At test time, both Teacher and Analyzer branches are stripped. Evaluated on five multi-turn VLM agent tasks across cognitive grid puzzles, 3D embodied control and navigation, and generative reasoning, PIVOT achieves 0.90 overall accuracy on Qwen2.5-VL-3B (+8% over SFT+GRPO baseline and +5% over previous SOTA) and scales to 0.92 on Qwen3-VL-2B (+12% over SFT+GRPO baseline). --- ### Narrowing the Horizon: Quantifying Topic Saliency Shifts in Generative Monoculture Authors: Oriane Peter, Elena Simperl, Kate Devlin Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35302v1 As Large Language Models (LLMs) become central to how we access and share information, they play an increasingly powerful role in shaping global knowledge. However, as these models evolve, their outputs risk converging into a \textit{generative monoculture}, where the diversity of perspectives they represent narrows over time. Studies at the model level often fail to pinpoint which specific topics or viewpoints are being marginalised or amplified in this process. In this paper, we introduce a method to measure shifts in topic saliency across model families, tracking what gains or loses prominence during post-training. Applying this approach to a case study of climate change discourse, we demonstrate how homogenisation affects the representation of diverse solutions across different models. We also test interventions to counter this trend, showing that specialised models can help preserve a broader range of perspectives. This underscores the importance of monitoring topic saliency to diagnose the risks of monoculture and to ensure AI systems reflect a pluralism of ideas. Data and Code are accessible \href{https://github.com/oriane/topic_saliency_shift}{here}. --- ### Training-Free Clinical Reasoning through Medical Ontologies and Cognitive Mapping: A Symbolic-Probabilistic Knowledge Graph Framework Authors: Surajit Das Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35298v1 Most clinical prediction systems learn patient-variable-outcome associations; we investigate a training-free diagnostic paradigm mapping patient observations to explicit medical knowledge. CKG Reasoner integrates candidate-specific Evidence Feature Nodes, patient-reference matching, a bounded Information Gate, knowledge-weighted evidence accumulation, disease similarity, and decisive clinical rules. Missing-aware normalization and coverage auditing distinguish absent from unavailable evidence. Candidate ranking is separate from outcome-label-independent K-means clustering, which uses four derived evidence coordinates (evidence strength, relative magnitude, directional similarity, and evidence completeness), not raw predictors or targets, to derive cohort-level assignments. Across six retrospective cohorts - four dengue (N = 1000, 1523, 989, 1018), malaria (N = 2190), and influenza (N = 4569) - a uniform, label-free, cohort-fitted K = 2 protocol yielded positive-class F1 scores of 0.996, 0.634, 0.936, 0.917, 0.695, and 0.842, and all-record accuracies of 0.996, 0.558, 0.914, 0.893, 0.707, and 0.906, respectively, with full partition-decision coverage using the frozen package and disease-specific knowledge representations. Neither scoring nor clustering uses outcome labels. Logistic regression provides a supervised baseline. Influenza incorporates confirmatory molecular PCR and is not independent pre-test prediction. Results characterize knowledge-grounded evidence separation, auditability, and sensitivity, not prospective clinical validity or comparative superiority. FOL/LLM-based clinical explanation remains unevaluated. --- ### LionMuon: Alternating Spectral and Sign Descent for Efficient Training Authors: Arman Bolatov, Artem Riabinin, Nikita Kornilov, Andrey Veprikov, Samuel Horváth, Martin Takáč, Aleksandr Beznosikov Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35297v1 Pretraining a language model takes enormous compute, and the right optimizer can save a good part of it. Muon's spectral step gives a stronger direction than a sign step, but it is expensive. Every step runs Newton-Schulz iterations on the full matrix and, in distributed training, an extra all-reduce. Sign steps, as in Lion and Signum, are cheap and stay local to each device. We propose LionMuon, which takes one Muon step every $P$ iterations and Lion steps in between, with a single dual-EMA momentum buffer shared by both. Muon's compute and communication are paid once per $P$ steps, and the optimizer state is half of AdamW's. A single-EMA variant, SignMuon, already improves on Muon. We prove complexity bounds under heavy-tailed noise in which the period sets an interpolation between Muon's and Lion's smoothness and noise constants, and which say when LionMuon is faster than both. On 124M and 355M models trained on FineWeb, LionMuon with $P=2$ and $P=5$ reaches a lower loss than Muon, AdamW, Lion and Signum at the same number of tokens. Under 4-GPU data-parallel training it reaches Muon's final loss with a third less wall-clock on PCIe, and it beats the communication-efficient Muon variants Dion and MuonBP on loss at no more exposed communication, while keeping the exact gradient. Code: https://github.com/brain-lab-research/lion-muon --- ### AbGaze: Attentive Geometric Representation Learning for End-to-End Antibody Design Authors: Jiashuo Wang, Siqi Fan, Yizhen Luo, Zaiqing Nie Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35296v1 Computational antibody design requires representations that capture the geometric patterns underlying antigen--antibody interactions, yet existing approaches often rely on scalar distances or surface-intrinsic features, leaving cross-molecular geometry largely implicit. We present AbGaze, an end-to-end antibody design framework based on attentive geometric representation learning, which encodes distance, spatial direction, and surface-normal orientation of antigen surfaces relative to antibody-residue local frames, and adaptively aggregates these geometric interactions according to their interfacial context. The learned interaction representation is shared across multi-CDR co-design, complex structure prediction, and affinity optimization, with local-frame geometric supervision further constraining the representation. AbGaze outperforms prior methods across all three tasks: relative to the second-best method, it improves amino-acid recovery by 7.1% and reduces structural error by 14.9% on average over the six CDRs, improves interface docking quality (DockQ) by 6.6%, and raises the affinity improvement rate (IMP) by 32.5%. --- ### Simulation-Based Inference for Plate Reverb System Identification Authors: Dylan Sechet, Marc Evrard, Matthieu Kowalski Published: 2026-09-28 Categories: eess.AS, cs.LG arXiv: https://arxiv.org/abs/2609.35295v1 We address Task A of the 1st DAFx Parameter Estimation Challenge, which aims to retrieve the physical parameters of a plate model from an impulse response. To do so, we use the Simulation-Based Inference (SBI) framework, in which we train a neural network to estimate a density over plate parameters given an impulse response, using a dataset generated by the simulator. Inference for a new impulse response then requires only a forward pass through the network, without involving the simulator. For each test observation, we fine-tune a specific network: additional simulation rounds are performed by sampling parameters from the current estimated distribution, simulating the corresponding impulse responses, and fine-tuning to produce the specialized network. --- ### Beyond Saying Less: Fine-Grained Alignment for Informative and Faithful Vision-Language Models Authors: Xingming Long, Jie Zhang, Yuecong Min, Shiguang Shan, Xilin Chen Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35294v1 Object hallucination remains a major challenge for large vision-language models. While off-policy preference optimization proves to be an effective solution, on-policy reinforcement learning provides a more promising direction as it directly targets a model's current failure modes. However, we find that without fine-grained reward formulation and allocation, on-policy optimization often falls into an easy shortcut: reducing hallucinations merely by saying less---making fewer valid claims. To comprehensively resolve this, we propose a fine-grained alignment framework that couples dense reward signals at the data level with precise credit assignment at the algorithmic level. Specifically, we first construct the Dense Object Presence and Absence (DOPA) dataset to address sparse annotations that prevent valid object claims from being verified and rewarded. DOPA exhaustively annotates the deterministic presence and absence of every concept across an expanded vocabulary, significantly increasing the density of reliable reward signals during on-policy rollouts. Second, we propose Subsentence-level Credit Assignment for on-Policy Optimization (SCAPO) to prevent response-level shared advantages from allowing local hallucinations to compromise all other valid outputs within the same response. By assigning credit to each subsentence independently based on its object claims, SCAPO can precisely reinforce faithful generations and penalize hallucinations. Furthermore, we leverage the resulting faithful image descriptions as auxiliary context to transfer generative gains to discriminative tasks. Experiments demonstrate that our method produces highly informative, faithful descriptions in generative tasks while yielding clear performance gains on discriminative evaluation. --- ### Decide, Don't Generate: Competitive Dimensional ABSA with Jev's Typed Decisions Authors: Yiqun Zhang, Peidong Wang, Zihan Wang, Shi Feng Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35293v1 Aspect-based sentiment analysis (ABSA) has largely turned to text generation. We show that competitive dimensional ABSA does not need it. Using Jev, a frozen model that answers typed questions with rubric scores, label probabilities, and yes/no judgments, we decompose all three tasks of SemEval-2026 Task III Track A into such decisions and align them with the annotation scheme through 488 coefficients fitted on CPU, with no text generation and no backbone tuning. On valence-arousal regression over ten corpora in six languages, the system reaches 1.0645 RMSE, the lowest aggregate error of any participating system. On triplet and quadruplet extraction, it reaches 52.09 and 44.06 continuous F1, above fine-tuned Llama-3.3-70B and GPT-OSS-120B baselines. Analyses and ablations show where the accuracy comes from: supervised calibration roughly halves the raw regression error, exact valence-arousal would add only 4.5 F1 to extraction, and the learned combination of span-boundary evidence, not any single signal, carries the extraction systems. --- ### Scaffold Then Internalize: Representation Injection for Diffusion Transformers Authors: Han Fu, Jiacheng Chen, Baoquan Zhao, Weidong Chen, Wei Liu, Qing Li, Xudong Mao Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.35292v1 Recent representation alignment (REPA) methods accelerate diffusion transformer training by aligning projections of the transformer's hidden states with representations from pretrained visual encoders. In this work, we explore a reverse and complementary direction to REPA: rather than projecting diffusion representations into the encoder's space, we inject encoder representations into the diffusion transformer, allowing them to actively participate in the denoising process. To this end, we introduce \textit{REPresentation Injection} (REPI), a training framework based on a scaffold-to-internalization strategy, in which projected encoder representations initially serve as a temporary scaffold and are then progressively internalized by the diffusion transformer. REPI outperforms REPA across a wide range of backbones and is highly complementary to it: combining the two yields substantial gains over either alone. Notably, with only 160K training steps, REPI + REPA matches vanilla SiT trained for 7M steps, a speedup of over $43.5\times$. Code will be available at https://jeneveuxpas.github.io/REPI --- ### Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment Authors: Shunchang Liu, Lukas Fluri, Xin Chen, Francesco Croce Published: 2026-09-28 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2609.35291v1 Modern AI models are aligned through post-training to adapt them to downstream tasks. Recent work shows that fine-tuning language models on narrow tasks can induce emergent misalignment (EM), causing broadly harmful behaviors beyond the training task. However, EM has been studied almost entirely in text-only tasks, leaving its manifestation in multimodal models unclear. In this paper, we define and analyze EM in the context of vision-language models. We first induce EM via fine-tuning on narrow multimodal tasks targeting vulnerable code, careless household-object use, and conspiratorial interpretations of ordinary scenes. Across fifteen commercial and open-source models with different scales, we find that narrow multimodal fine-tuning can induce coherent and broadly misaligned behavior that transfers to unrelated tasks, including misaligned opinions, visual factual dishonesty, unsafe image generation, vulnerability to visual jailbreaks, and risky agentic actions. We further find that multimodal EM does not depend on the apparent harmfulness of training data but is sensitive to training-evaluation modality alignment. EM can arise under both supervised fine-tuning and preference optimization and can propagate through intermediate reasoning. Finally, we explore several mitigation strategies, including prompt inoculation, benign continued training, and activation-level steering, which can partially reduce EM. Overall, our findings suggest that multimodal EM reflects a behavioral shift rather than a general loss of capability, extending beyond text to the visual modality. --- ### EvoIn: Bridging Evolution and Internalization for Agent Fine-Tuning Authors: Shihan Dou, Shaofan Liu, Zhonghang Lu, Jiahang Lin, Shichun Liu, Binghai Wang, Jiajie Jin, Guanting Dong, Tao Gui, Qi Zhang, Xuanjing Huang Published: 2026-09-28 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.35290v1 Recent work has explored improving agents by jointly evolving their harnesses and models, but often takes a ''potpourri'' approach that bundles together new tools, new decision-making procedures, and model adaptation to the evolved harness under a single notion of agent improvement. In this paper, we instead investigate how agents can improve their decision-making procedures. In particular, we propose EvoIn, an agent fine-tuning framework that bridges evolution and internalization. EvoIn first analyzes agent execution traces to evolve and validate new decision-making procedures by temporarily instantiating them in the harness. The validated procedures guide the agent to generate improved reasoning traces. These traces are then rewritten into self-contained reasoning traces, removing explicit references to harness instructions while expressing the induced decision logic as the model's own reasoning. Finally, EvoIn fine-tunes the model on the rewritten traces, internalizing these procedures so that the improved decision-making persists without the evolved harness at inference time. We evaluate EvoIn on diverse benchmarks and find that it consistently enables agents to learn stronger decision-making procedures, raising the pass rate by 10.9 points in-domain and by 9.2 points out-of-domain. Results further show that the internalized decision procedures generalize to unseen tasks. Case studies show that agents can learn to decide how to solve a task before solving it, for example by checking a document's length to choose between reading it in full and searching it. EvoIn is also broadly applicable, showing consistent improvements on another model family. --- ### Domain-adaptive Zero-Shot Image Enhancement via Locality-Constrained Diffusion Guidance Authors: Theresa Neubauer, Dimitrios Lenis, Astrid Berg, Maria Wimmer, Gaia Romana De Paolis, Philip Matthias Winter, David Major, Johannes Novotny, Ariharasudhan Muthusami, Katja Bühler Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35289v1 Denoising Diffusion Probabilistic Models have shown remarkable performance in unconditional image generation. In order to generate images with desired semantics, recent works have restricted the solution space by using guidance constraints in the diffusion sampling process. However, for image enhancement across different domains, these methods struggle to balance two main requirements: looking realistic in the target domain (photorealistic images) and preserving relevant features of the source domain, e.g., low-quality renderings or art paintings. Here, small local changes can alter the fidelity of the image completely, while large changes in other regions might be insignificant. We introduce LocDiff, a locality-constrained guidance method for image enhancement, which serves as a zero-shot extension to pre-trained diffusion models, ensuring the preservation of critical features during domain adaptation. In this way, we retain important local features, while allowing less critical regions to remain unconstrained and not interfere with the guidance process for relevant regions. We evaluate our method on two different domain-shift tasks: For art-to-photo translation, we apply the method in a fully zero-shot setting, preserving facial identity from paintings while generating photorealistic details. For enhancing low-quality fetal ultrasound renderings, we demonstrate zero-shot inference with auxiliary prior alignment. Here, the objective is to artificially add high-resolution characteristics and produce photorealistic ultrasound renderings, a target domain for which no ground truth distribution exists. Our experimental results demonstrate that LocDiff achieves favorable realism-faithfulness trade-offs compared to state-of-the-art methods, enabling controllable cross-domain enhancement. --- ### $λ$-JEPA Spectral Anti-Collapse Regularization for Self-Supervised Learning Authors: Berker Demirel, Clémentine Dominé, Valentino Maiorca, Marco Fumero, Marco Mondelli, Francesco Locatello Published: 2026-09-28 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2609.35288v2 Joint-embedding self-supervised learning typically combines an invariance objective across augmented views with additional mechanisms to prevent representational collapse. These objectives are often applied after a projection head, while downstream tasks use the backbone representation before the projector. We find that this mismatch does not necessarily prevent dimensional collapse in the backbone, which can retain low effective rank and potentially limit downstream transfer. To address this, we introduce SACReg, a spectral anti-collapse regularizer motivated by an analysis of $λ$-balance, which captures the relative scale of weight matrices across layers. In a two-layer linear network, we show that (i) $λ$-balance prevents collapse, and (ii) our regularizer applied to the backbone induces $λ$-balance. In the nonlinear case, this regularizer leads to anti-collapse as well and, in realistic architectures on ImageNet100, it empirically increases the representations' ranks. We apply SACReg to JEPA and propose $λ$-JEPA, which improves over LeJEPA and VISReg on ImageNet-1k classification and in average linear-probe transfer performance across eight downstream image datasets. On video self-supervised learning, $λ$-JEPA improves over LeVJEPA and V-JEPA 2 on the Something-Something-v2 and Kinetics-400 benchmarks. Code is available at https://github.com/berkerdemirel/lambda-jepa. --- ### The Argument and the Letterhead: Source-Position Coherence in AI Evaluation Authors: Michele Loi Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35286v1 An argument can be surprising coming from a particular speaker without being a bad argument. Do AI evaluators keep these judgments apart? Two preregistered descriptive studies and a later Jev supplement collected 2,976 usable evaluations of six fixed texts about US AI policy, Germany's debt brake and Swiss nuclear energy. Each text was presented under several source attributions. The key comparison asks whether the gap between two sources changes when the argument changes. On Sol, for example, a national-security argument received mean ratings of 0.359 under CODEPINK and 0.639 under College Republicans; a civil-rights argument received 0.742 and 0.721. A constant preference for one source cannot explain that pattern. Related interactions appeared across topics and recent model configurations, including those with reasoning enabled, while several comparisons yielded small effects. The later European Jev supplement yielded five interactions below the adopted absolute reference of 0.05; its distinct rubric and interrupted collection limit comparison with the chat systems. Some written evaluations explicitly invoked a mismatch between a source and its attributed position. Taken together, the numerical and verbal evidence supports source-position coherence as a plausible explanation, alongside competing accounts involving credibility, authenticity and interpretation of the task. The paper develops this inference through controlled comparisons, reports conditional post hoc p-values in an appendix, and documents the human decisions and delegated checks behind an AI-conducted study. --- ### Textual User Taste: Natural-Language User Context for Foundation-Model Recommender System at Scale Authors: Ghazal Fazelnia, Paul Gigioli, Eliza Klyce, Sharon Zheng, Katie Zelvin, Ye Myat Thein, Anurag Deshpande, Seda Davtyan, Kate Remeika, Maya Hristakeva, Erik Franco, Karen Banzon et al. Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35285v1 Foundation model recommender systems require user context that can be consumed by large language models, reasoned over, and refined through natural-language interaction. Traditional behavioral embedding vectors remain highly effective for retrieval and ranking, but they are opaque to users and not natively expressed for language model workflows. We present Textual User Taste, a system that generates structured natural-language taste profiles from listening behavior, interaction signals, content metadata, and optional user feedback, and deploys them to millions of Spotify users. We describe the end-to-end production lifecycle required to generate, evaluate, optimize, and maintain these representations at industrial scale, including prompt development and compression, user steering, and integration with downstream personalization systems. Because no unique ground-truth taste profile exists, we introduce a multi-faceted evaluation framework to evaluate taste profiles as a production representation: they carry user-specific predictive signal independently, and when integrated with behavioral embeddings, improve MRR by 0.6% for future-track prediction and NDCG@7 by 2.2% for search ranking. Our evaluation also reveals that taste profiles support positive natural-language steering, while exposing important limitations, including challenges with negation and short-term temporal adaptation. These findings position taste profiles not as replacements for behavioral embeddings, but as an interpretable and steerable interface between evolving user context and foundation-model recommender systems. --- ### Measuring Collapse and Correction in Homogeneous-Panel LLM Debate Authors: Xin Li, Mengbing Liu, Chau Yuen Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35279v1 Multi-agent large language model (LLM) debate is often evaluated by whether final answers improve, but movement is not necessarily improvement: the same discussion can rescue an initially wrong majority or destroy an initially correct one. Standard final-accuracy evaluations conflate these opposing mechanisms. We introduce an auditable protocol for homogeneous debate on multiple-choice questions (MCQs) that records each run as a transition ledger over collapse, correction, onset, and signed intervention utility. On 6,925 MMLU-Pro debates, the protocol identifies 253 collapses and a parallel correction ledger that changes how interventions should be judged. Replay experiments reveal the central tradeoff: a leave-one-model-out probe-gated freeze prevents 29 collapses but loses 108 corrections under equal weights, so collapse prevention alone can recommend the wrong policy. A compact pre-debate 8-probe screen is a triage signal: its unadjusted family-level association with conditional-collapse risk is high (G=7, Spearman rho=0.893, exact two-sided p=0.0123), but initial-majority accuracy is a close comparator (rho=0.821; family partial rho=0.767, p=0.0877), so we do not treat it as calibrated or capability-adjusted prediction. Round-level traces localize many collapses to the first debate round, where early disagreement can precede both harmful cascades and useful recovery. We release replayable schemas, coders, audits, cost cards, and zero-API rebuild scripts so future model-scaffold rows can be compared under the same denominators and signed utility ledger. --- ### Evaluating Hierarchy-Aware Deep Learning for the Recognition of Tironian Notes Authors: Yule Kang, Thomas Gorges, Janne van der Loop, Franziska Marske, Nikolaus Weichselbaumer, Tino Licht, Vincent Christlein Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35277v1 Tironian notes are generally regarded as the first Latin shorthand system and are notable for their large, fine-grained symbol inventory. Their high visual similarity and large class set make manual reading time-consuming, leaving manuscripts that contain Tironian notes inaccessible to many researchers. Automatic recognition is also challenging because models must distinguish subtle differences in stroke shape and sign structure while realistic training data remain scarce. However, standard flat classifiers do not explicitly use visual or structural relations between related signs. This paper investigates whether structural relationships between Tironian notes can support automatic recognition. We use the Supertextus Notarum Tironianarum (SNT) by Martin Hellmann, which provides idealized sign forms and a hierarchical organization of Tironian notes. We compare flat ResNet18, ConvNeXt, Shifted Window Transformer (Swin), and Vision Transformer (ViT) classifiers with Hierarchical Deep Convolutional Neural Network (HD-CNN)-style coarse-to-fine models and hierarchy-aware routing models based on visual class cleaning and similarity-based re-clustering. The models are evaluated on handwritten samples and manuscript-domain samples from Vergilius Turonensis, both with and without limited few-shot adaptation to the manuscript domain. The results show that the relative performance of flat and hierarchical models depends on adaptation. On Vergilius Turonensis, HD-CNN achieves the best non-adapted Top-1 result with 45.43%, while flat classification reaches the best Top-1 result after few-shot adaptation with 82.09%. Overall, the results indicate that hierarchical structure can support Tironian note recognition, especially under non-adapted conditions. --- ### Multi-Attractor GNNs: Set-Valued Expressivity Beyond Unique Equilibria Authors: Jialin Liu Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.35274v1 Recurrent and equilibrium graph neural networks (GNNs) often enforce a unique fixed point or use one training target per graph. Yet many combinatorial and scientific problems admit multiple valid solutions, with no preferred one. A designated target can then impose an arbitrary selection rule. For tasks invariant to node relabeling, a symmetric graph may have a symmetric solution set but no symmetric solution. We show that multiple equilibria enable one weight-tied message-passing GNN to represent set-valued equivariant maps: different initializations approach different valid solutions. Under stated regularity assumptions, we first construct globally Lipschitz, permutation-equivariant dynamics that converge almost surely to valid solutions and reach every solution branch with positive probability. We then establish approximate realization by recurrent message passing with continuous component maps, with arbitrarily small update and limiting errors and arbitrarily high probability. This goes beyond standard universality arguments: although message passing alone cannot distinguish symmetric nodes, the evolving state keeps nodes distinguishable at every finite step without auxiliary node identifiers. Such dynamics can be learned without solution labels using problem-specific energies. On Ising ground states, structural module detection in protein graphs, and chemical reaction steady states, the learned updates produce multiple high-quality predictions with high numerical convergence rates. They achieve better average solution quality than the tested unique-equilibrium, single-target, and feedforward baselines, while remaining competitive with much larger diffusion-based solvers. --- ### When Words Speak Louder than Images: Towards Understanding Language Bias in Vision-Language Models Authors: Yizhou Fang, Siyue Chen, Zimo Qi, Zhiyu Xue, Xi Chen, Guangliang Liu Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35272v1 Despite substantial progress across downstream applications, vision-language models (VLMs) remain susceptible to language bias, often prioritizing linguistic cues over visual evidence and consequently producing incorrect predictions. Prior studies have proposed various approaches to understanding and mitigating language bias in VLMs, yet their findings often conflict due to the difficulty of tracing how language bias propagates within black-box VLMs. Building on the word completion task, we trace how language bias propagates through VLM inference by (1) proposing a diagnostic framework that decomposes the inference process into four distinct yet interdependent stages to trace the propagation of language bias; and (2) examining how two key factors underlying language bias, i.e., linguistic priors and cross-modal coverage, evolve across these stages and ultimately give rise to incorrect predictions. The linguistic prior captures the strength of statistical bias induced by the language model component of a VLM and represents the origin of language bias, whereas cross-modal coverage measures the extent to which linguistic cues cover the visual content. By decomposing inference into four stages and characterizing the interplay between linguistic priors and cross-modal coverage across these stages, we propose a systematic framework for tracing the propagation of language bias throughout the inference process; and uncover the underlying mechanism of language bias by revealing the interplay between linguistic priors and cross-modal coverage. --- ### eval-unlearn: Benchmarking unlearning in Text-to-Image Diffusion Models Authors: Mansi, Nikhil Raghavan, Zixia Huang, Kai Sheng Ong, Ji Shen Lim, Brandon Siao Xiang Ling, Francesco Leofante Published: 2026-09-28 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.35269v1 The rising number of concept unlearning techniques for text-to-image (T2I) diffusion models has produced a fragmented evaluation landscape. Methods are assessed under heterogeneous experimental conditions making principled cross-method comparison difficult. We present eval-unlearn, an open-source Python library providing a unified, reproducible benchmarking framework for concept unlearning in T2I Diffusion models. eval-unlearn integrates twelve published unlearning techniques spanning fine-tuning, closed-form model editing, and inference-time intervention, alongside nine complementary evaluation metrics covering erasure efficacy, adversarial robustness, generative quality, and concept retention. Its plugin architecture lets third-party techniques and metrics self-register without modifying the core framework, and its streaming, batched pipeline supports efficient evaluation of both standard NSFW concepts and arbitrary general concepts. As a further contribution, we release a public leaderboard on HuggingFace along with an interactive tool for real-time evaluation of unlearning techniques. The leaderboard compares nudity concept erasure case study across all twelve techniques, exposing significant accuracy-quality trade-offs that are obscured by heterogeneous evaluation. eval-unlearn is released under the MIT license; the package, code, leaderboard, and documentation are all available at https://eval-unlearn.readthedocs.io. --- ### SpikeCredit: Temporal Credit Carrier for Reinforcement Learning with Sparse Rewards Authors: Yingchao Yu, Pengfei Sun, Wenxuan Pan, Wei Chen, Yitian Hong, Kuangrong Hao, Yaochu Jin Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35268v1 Reinforcement learning (RL) with sparse rewards is challenging because delayed outcomes provide little guidance about which intermediate computations caused success or failure. We argue that reliable credit assignment requires policy dynamics that preserve and expose credit-relevant information over time, a role we formalize as Temporal Credit Carriers (TCCs) and that spiking neural networks (SNNs) naturally fulfill through graded membrane traces and event-driven spikes. Based on this hypothesis, we propose SpikeCredit, an SNN-based framework for RL with sparse rewards that first performs task-adaptive TCC selection and then closes the loop between a fast TCC-reading pathway, where self-motion feedback constraint uses local behavior-grounded cues to constrain transition-level credit recovery, and a slow TCC-writing pathway, where credit-targeted trace alignment feeds recovered credit back into the actor to make future TCC dynamics more credit-readable. Across sparse-reward MuJoCo tasks, SpikeCredit improves Last10 return over sparse SNN baselines by +1169% on Ant, +953% on Hopper, +723% on Swimmer, and +1781% on Walker2d, and exceeds the dense-reward baseline on Swimmer by +113%. Mechanistic analyses further show substantially stronger alignment with dense rewards than the sparse SNN baseline. These results position spiking dynamics as credit-preserving substrates for sparse-reward RL. --- ### WavePP: High-Throughput Pipeline Parallel LLM Prefill under Prefix Reuse Authors: Aaryam Sharma Published: 2026-09-28 Categories: cs.DC, cs.AI arXiv: https://arxiv.org/abs/2609.35263v1 Pipeline parallelism can improve prefill throughput by processing multiple request chunks concurrently across different stages of the model. However, keeping the pipeline fully utilized requires efficient scheduling and request preparation. In systems where stages retain and evict cache state independently, a local cache hit does not guarantee that the same prefix can be reused across the pipeline. Here, coordination overhead can impede request admission cadence and thus reduce overall throughput. In this paper, we present WavePP, a prefill runtime built on top of TensorRT-LLM that addresses these challenges by overlapping request admission with pipeline execution. WavePP asynchronously finds a prefix that can be reused across all stages, protects the cached state, and reserves space for the remaining input while earlier requests continue to execute. It subsequently plans the chunk sizes of each request dynamically to maximize pipeline fill. Each stage then completes the local preparation before executing the request. In the same system and pipeline topology, WavePP improves TensorRT-LLM's prefill throughput in 37 of 40 tested settings on GLM 5.2 and MiniMax M2.7. At concurrency 128 with high cache reuse, these changes increase throughput by factors of 2.91 and 2.02, respectively. Across 28 Kimi K3 settings, WavePP also has the highest measured throughput in all 18 settings at concurrency eight or higher, compared with tensor/expert-parallel and pipeline-parallel baselines from TRT-LLM, SGLang, and vLLM. --- ### Rubric-Aware On-Policy Self-Distillation for LLM Personalization Authors: Yilun Qiu, Xiaoyan Zhao, Chengbing Wang, Cilin Yan, Rui Zu, Wanyang Zhang, Xiaolong Jiang, Jiayin Cai, Yang Zhang Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35262v1 LLM personalization aims to generate responses aligned with individual users' preferences and needs. User-specific rubrics make these expectations explicit, providing direct supervision on what a satisfactory answer should cover. Existing rubric-guided approaches, however, exploit such guidance only at a coarse granularity, either by using rubrics to supervise the prediction of relevant aspects for subsequent generation or by reducing aspect coverage to a single response-level reward for reinforcement learning. This leaves a gap between specifying what a personalized answer should contain and teaching the model how to generate it. To bridge this gap, we propose GRASP, a rubric-aware on-policy self-distillation framework for LLM personalization that turns user-specific rubric aspects into fine-grained, token-level supervision. Specifically, GRASP pairs a rubric-free student with a rubric-informed teacher that additionally receives the target user-specific rubrics. By aligning their next-token distributions along on-policy trajectories generated by the student, GRASP transfers the teacher's rubric-conditioned guidance into the student, translating user-specific semantic requirements into dense token-level supervision. Since rubric-informed teachers can still produce inadequate supervision, we further introduce Rubric-based Teacher Validation (RTV), which retains only instances where the teacher sufficiently covers the target aspects, improving both supervision quality and training efficiency. Experiments on the LaMP-QA benchmark for personalized question answering demonstrate that GRASP achieves state-of-the-art performance across multiple backbones, supporting the effectiveness of rubric-guided token-level supervision for personalization. To ensure reproducibility, our code is available at https://github.com/SnowCharmQ/GRASP. --- ### Imprint Reader: From Weight-Update Readout to Behavioral Intervention Authors: Guanxu Chen, Qihao Lin, Jing Shao Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35261v1 As language models take a growing role in AI development, a natural aspiration is for them to reflect on their own learning process, as humans do, and use that reflection to improve themselves. At the same time, these models have an advantage that human learners lack, since training leaves parameter-level traces that can, in principle, be inspected directly. However, current models cannot decode these traces into an explicit account of what they have learned. To this end, we introduce the \textit{Imprint Reader}, a model trained with \textit{Semantic Mount-and-Read Tuning} (SaRT) to describe frozen weight updates. SMaRT mounts each update onto the Reader and uses an anchor-free meta-query to elicit a natural-language description, while no-change and random-perturbation controls discourage unsupported claims. On held-out updates, the joint Reader reaches judge-based Pass@100 of $2\%$ for knowledge and $16\%$ for behavior. These results demonstrate the feasibility of natural-language readout while pointing to reliability across updates as the next step. Beyond free-form generation, the Reader provides a differentiable proxy for the gap between a specified target behavior and a candidate weight update. Its coordinate-aligned gradients support intervention through MetaEdit. At a $0.5\%$ pruning rate, Reader-guided selection raises measured harmful-prompt refusal from $57.9\%$ to $64.1\%$ under a safety-maintenance target. Using behavior descriptions without target-task training data, MetaEdit increases the frequency of backtracking and sub-goal expressions in mathematical reasoning traces and raises BFCL Overall from $41.69\%$ to $44.60\%$. --- ### Latency and accuracy tradeoffs in Spiking Neural Networks Authors: Zhanglu Yan, Zixuan Zhu, Kaiwen Tang, Yuyang Cai, Qianhui Liu, Weng-Fai Wong Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35260v1 Spiking neural networks are attractive for low-power speech command recognition, yet their latency has received far less attention than their energy efficiency, and their multi-timestep execution is widely assumed to make them slower than quantized neural networks. This paper challenges the assumption that more local timesteps necessarily imply higher network latency. By overlapping computation across adjacent layers at the timestep level, SNNs may complete execution in less time than comparable bit-serial QNNs. However, this overlap relies on spikes firing on incomplete inputs, and a spike once generated cannot be withdrawn, so its error persists and reduces accuracy. Waiting for more input before firing would seem to improve accuracy at the cost of reduced overlap. Yet we find and prove that this intuition fails at some layers, where even a small increase in waiting can change spike timing and downstream computation, making the network both slower and less accurate. We therefore propose a Pipeline Delay Search method which selects each layer's delay by balancing task-level accuracy gains against added network latency. We then adapt the selected configurations through spike-based quantization-aware training and bounded tuning of firing thresholds and initial membrane potentials. Together, these steps form Falcon, a framework for Fine-grained Analysis of Latency and Controlled firing which systematically analyzes and optimizes SNN latency under a spatial analog compute-in-memory mapping with shared digital engines. We evaluate Falcon on GSCV2 and SSC, achieving competitive accuracies of 96.31 and 83.02 at modeled network-core latencies of 119.64 and 124.00us, respectively. Together, our analysis and results show that SNNs can compute more yet finish faster, and wait longer yet predict worse, highlighting why Falcon matters for both latency and accuracy. --- ### On-Policy or Off-Policy Learning? A Systematic Study of Distillation Dynamics Authors: Julianna Piskorz, Antonin Berthon, Mihaela van der Schaar Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35259v1 On-policy learning has been argued to reduce catastrophic forgetting, produce sparser parameter updates, and improve generalisation. However, existing comparisons between supervised fine-tuning and reinforcement learning vary many factors simultaneously, making the contribution of rollout policy difficult to isolate. We study the effect of rollout policy in a controlled strong-to-weak distillation setting, by independently varying rollout policy, token-level KL direction, and learning rate across the Llama3 and Qwen2.5 model families and reasoning tasks spanning scientific, medical, and arithmetic domains. Our analysis reveals a nuanced picture of distillation dynamics in which rollout policy does not necessarily play a central role. Instead, token-level KL direction more clearly shapes task performance and output coverage, while learning rate governs forgetting and update sparsity. Analysis of KL gradients and experiments along a continuous student-teacher rollout-policy spectrum explain this pattern: forward KL is remarkably robust to rollout policy, with its performance stable and strong despite changes to the rollout policy, whereas reverse KL is substantially more sensitive and favours student-generated rollouts. On-policy data nevertheless improves generalisation to harder variants of the Countdown arithmetic task under both KL directions, although this advantage does not reliably persist after subsequent RLVR. Our broader conclusions remain robust to removing gradient clipping, using sampled KL estimators, and training on tasks requiring longer reasoning chains. Overall, our results challenge the view that on-policy rollouts are inherently preferable and show that their value depends critically on the objective, evaluation setting, and optimisation hyperparameters. --- ### AIM-ZO: Activation-Informed Subspace Maintenance for Zeroth-Order LLM Fine-Tuning Authors: Yue Xie, Zhi Zheng, Yunpeng Ba, Xuyang Wu, Xialiang Tong, Zhichao Lu, Tao Zhong, Zhenkun Wang Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35257v1 Zeroth-order (ZO) optimization offers a memory-efficient alternative for LLM fine-tuning by estimating updates only from forward evaluations of perturbed parameters, without backpropagation or activation storage. However, in billion-parameter LLMs, isotropic perturbations often waste many forward evaluations on weakly informative directions. To make these evaluations more informative, existing ZO methods restrict perturbations to low-dimensional subspaces. Yet the quality of these subspaces is critical: overly compressed or poorly maintained spaces can miss useful update directions. To obtain a high-quality subspace for ZO updates, this paper proposes AIM-ZO, a ZO fine-tuning method based on Activation-Informed Subspace Maintenance. AIM-ZO uses forward activations as local directional information and continuously integrates them into a broad, evolving subspace over training. To access broader gradient-relevant structure while keeping individual perturbations low-dimensional, AIM-ZO activates only a smaller set of shared and sampled directions, decoupling the maintained width from the active width. We evaluate AIM-ZO across 5 LLMs and 11 downstream tasks under matched forward-evaluation budgets; its six-task average exceeds the strongest fully evaluated ZO baseline by 1.26 percentage points on OPT-2.7B and MeZO by 2.85 percentage points on OPT-30B. Our code is available at https://github.com/EkkoXy/AIM-ZO --- ### Towards Reliable AI Data Scientists: Data Agents with Workflow Harnesses Authors: Huachi Zhou, Yujing Zhang, Jiahe Du, Jiacheng Cai, Zijin Hong, Chuang Zhou, Zheng Yuan, Qinggang Zhang, Qing Li, Xiao Huang Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35255v1 Large language model agents are increasingly deployed for data-intensive work, yet reliable data analysis requires more than general-purpose reasoning and ad hoc tool augmentation. Data Agents, equipped with workflow harnesses, offer a promising paradigm for automating the end-to-end data science lifecycle. This paper examines Data Agents from a harness-centric perspective. First, we introduce a taxonomy of Data Agents and associated data environments, organizing the literature around five functional stages: perception, planning, execution, verification, and repair. Second, we analyze the key technical routes within each stage, identifying 15 distinct approaches ranging from data structure probing to data state reconstruction. Third, we identify four open reliability problems: inactive semantic calibration, missing clarification, missing experience transfer, and the missing verification-repair repository. These problems explain why silent failures can persist even when individual components function correctly, highlighting the need for rigorous workflow harnesses and shared reliability resources. Finally, we summarize the horizontal task families of Data Agents, examine their vertical application settings, and benchmarks for evaluation, while maintaining a companion repository at https://github.com/DEEP-PolyU/Awesome-Data-Agents. --- ### SCBO: Semantically Coherent Batching and Ordering for LLM-Based Social Surveys Authors: Yuanzi Li, Lingjie Wang, Zihang Tian, Lei Wang, Xu Chen Published: 2026-09-28 Categories: cs.CL, cs.CY arXiv: https://arxiv.org/abs/2609.35250v1 Large Language Models (LLMs) offer a scalable way to simulate survey respondents using demographic profiles and observed reference responses. However, the conventional approach of predicting one question per prompt repeatedly encodes the same context, limits each target to a narrow set of reference responses, and prevents later predictions from using information in earlier answers. Predicting multiple questions in one prompt can reduce these costs, share a broader pool of references, and let later predictions build on earlier ones. This requires forming coherent batches, selecting shared references, and ordering questions and references effectively. We propose Semantically Coherent Batching and Ordering (SCBO), a training-free framework that addresses these challenges. SCBO first uses an LLM to extract compact semantic representations from survey items and filter out template noise. It then groups related questions into batches and builds a shared reference bank using target-specific retrieval and centroid-based completion. Finally, it orders target questions from easy to hard and arranges references according to their semantic alignment with those questions. Experiments on four large-scale survey datasets and four LLMs show that SCBO substantially reduces token consumption and inference time while generally improving prediction accuracy over a non-batched baseline. Code is available at https://anonymous.4open.science/r/SCBO-41D8. --- ### Spatial Grafting: Grounding 3D Features for Flow-Matching Robot Policies Authors: Dingsheng Liu, Yangzheng Wu, Mahboubeh Asadi, Zhiyuan Li, Jinbang Huang, Yixin Xiao, Tongtong Cao, Yingxue Zhang Published: 2026-09-28 Categories: cs.RO, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.35249v1 Pretrained robot manipulation policies such as vision-language-action models (VLAs) or world-action models (WAMs) leave interaction-relevant metric geometry implicit. Recent breakthroughs in spatial reconstruction can supply the necessary geometry reliably, but their features describe local shape without stating where it lies with respect to the robot. How best to deliver these features to a pretrained policy remains unresolved. We propose Spatial Grafting, a versatile, lightweight spatial module that binds frozen reconstruction features to metric, robot-relative geometry. Spatial Grafting constructs metric-grounded spatial tokens and injects them into the flow-matching action expert through cross-attention, without modifying the host's perceptual pathway, so the host retains the full benefit of its pretraining. We evaluate it more broadly than any geometry-aware policy we compare against: one graft architecture, with no per-host redesign, on two VLAs and two WAMs, across four simulation benchmarks that span short-horizon manipulation, visual robustness, clutter and long-horizon mobile manipulation, and on three real-robot platforms with single- and dual-arm configurations. On RoboTwin 2.0, a dual-arm manipulation benchmark, the graft improves every host across VLAs and WAMs. Grafted $π_{0.5}$ gains 11.3% and 15.6% on clean and randomized scenes, reaching 94.0% and 92.4%, above the strongest published 3D-conditioned policy, WAM4D (93.8% and 89.9%). The margin widens as the horizon lengthens: on tasks from BEHAVIOR-1K, a dual-arm mobile manipulation challenge scored by average task progress, it surpasses the 2025 challenge winner on five of six tasks,by up to 0.47 Q-score, and exceeds a map-conditioned spatial policy on average across the three tasks both report. --- ### AnswerMap: Faithful Spatial Interpretability of VLMs from Answer Posteriors Authors: Mohamed Eltahir, Fardows Adam, Duaa M. Tahir, Lama Alamoudi, Sana Ammar, Atheer A. Alboloshi, Jory Albluey, Tanveer Hussain, Naeemullah Khan Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.35247v1 When a VLM answers a visual query, current interpretability tools rely on text rationales, which use a mismatched modality, or on internal read-outs, which originate too early to reflect the final output and require white-box access to the model. We introduce AnswerMap, a training-free, task-agnostic, black-box visual rationale constructed from the output head. The image is cut into K row and K column bands, each shown alone to the frozen model along with the query in the format of a yes/no relevance question. The outer product of the row and column ``yes'' posteriors gives the query-conditioned spatial map. Crucially, by defining a fixed read-out R (e.g., expectation, maximum) on top of AnswerMap, we can derive continuous outputs like location natively. This bypasses the reliance on discrete text tokens for continuous-output tasks and guarantees an image-dependent answer by construction. However, a rationale can be confabulated, so we validate AnswerMap across four models and three query distributions with two tests: (a) agreement with the model's own generated point and (b) deletion of the map's region. The map lands where the model points (AUC 0.85 against 0.38 for attention), and deleting its region flips 53% of correct answers (against 19% for attention's). Beyond establishing faithfulness, we demonstrate the map's task-agnostic utility through three distinct read-outs: its maximum flags hallucinated objects without generation, its expectation localizes correctly when the model's own pointing fails, and its top-mass region, fed back as a crop, fixes half of the model's wrong answers. AnswerMap thus offers a new lens on VLM interpretability and, through its read-outs, a new output interface for visual tasks beyond text tokens. --- ### DrawingsDreamer: A Unified Multi-View Engineering Drawings Generation Model Authors: Shurui Liu, Weide Chen, Changwang Yi, Ancong Wu Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35242v1 Scalable Vector Graphics (SVG) are essential for modern industrial Computer-Aided Design (CAD). However, existing autoregressive SVG generation models are predominantly tailored for artistic creation and struggle to maintain the rigorous geometric fidelity and cross-view spatial alignment required for engineering drawings. To bridge this gap, we introduce \textbf{DrawingsDreamer}, a unified Large Language Model (LLM)-driven framework for multi-view vector-based engineering drawings generation. By formulating the generation of multi-view engineering drawings purely as a sequence modeling task, we eliminate the need of raster image encoders. We propose a Streamlined Representation utilizing hierarchical postfix tokenization, which guides the model to establish local geometric coordinates before assigning semantic boundaries. Optimized via a progressive task-aware curriculum schedule, \textbf{DrawingsDreamer} effectively transitions from localized structural repair to macroscopic generation in a unified model. Extensive experiments demonstrate that our unified model achieves strong performance in both geometric fidelity and syntactic accuracy across diverse conditional and unconditional generation tasks. --- ### Disentangling Lung-Cancer CT/LDCT AI: A Systematic Evidence Map of Clinical Tasks, Evidence Chains, and Translational Gaps Authors: Surajit Das Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35240v1 Artificial-intelligence studies using computed tomography (CT) for lung cancer are often broadly labelled "prediction" despite addressing clinically distinct tasks. We systematically mapped CT/low-dose CT (LDCT)-centered lung-cancer AI using five-database retrieval, full-text eligibility assessment, role-aware modality/omics extraction, clinical-task classification, and a Multi-Tier Evidence Graph (MTEG). The final corpus comprised 293 studies (2016-2026): 230 Detection, 8 future Risk-prediction, and 55 Other studies. Clinical variables (96.2%), 3D CT/LDCT (73.0%), and radiomics (63.5%) predominated, whereas external validation (29.0%), calibration (20.5%), decision-curve analysis (13.0%), longitudinal CT (17.7%), and saliency/attribution XAI (21.5%) were less frequent. The MTEG comprised 377 nodes and 3,444 edges; only 31 studies (10.6%) completed the six-tier substantive evidence chain, with greatest attrition at reasoning/explanation. Overall, the literature is detection-dominated, genuine future risk prediction remains uncommon, and complete translational evidence chains are rare. --- ### Long-Horizon Scaling: How Model Capabilities Shape the Returns to Computation Authors: Haoyu Zheng, Zhengyu Chen, Huaisheng Zhu, Ruishan Fang, Teng Xiao, Yiwei Li, Jingang Wang, Wenqiao Zhang Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35236v1 Long-horizon agents improve solutions through sustained interaction, execution, and task feedback. Scaling studies relate performance to resources and capabilities, yet how existing capabilities shape returns to extended interaction remains less understood. To address this gap, we analyze AutoLab and EdgeBench, two long-horizon benchmarks. We find that starting performance and subsequent growth are associated with different capabilities: within a task category, similar early scores can precede different later gains. To formalize this finding, we model capability-time scaling with category-specific logistic power laws shared across models. Fitted to early trajectories, these curves extrapolate the observed models' category-average scores to later computation. However, rising average scores mask narrowing improvement opportunities: later gains concentrate among fewer improving models. High final scores and continued improvement also have distinct capability profiles. Predicted mean gains estimate each model's fraction of improving tasks; averaging these estimates forecasts the average share of improving models. These uneven returns motivate deciding whether a specific run should continue. We therefore derive a continuation policy to save time and compute with limited score loss. The policy conditions growth predictions on the run's observed progress and weighs immediate and delayed gains against computation costs. In replay with training and price calibration based on other models' histories, the policy saves roughly one-third of full-run time, with relative score losses of 2.4% on AutoLab individual runs and 3.3% on EdgeBench published mean curves. Our repository is available at https://github.com/Chihaya-Anon-chan/long-horizon-scaling. --- ### EP-Mem: Elastic Privacy Memory for Social Relationship-Aware LLM Agents Authors: Fengzhou Sun, Yuan Zhang, Xintong Yu, Jinyao Yan Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35233v1 Large language model (LLM) agents face critical privacy risks when acting as delegates in human-agent-human communication. To prevent such breaches, agents must understand users' social relationships and adhere to context-dependent social information disclosure boundaries. Current studies on agent memory privacy focus on instantaneous interactions, leaving the long-term relational disclosure problem unexplored. In this paper, we propose EP-Mem, an Elastic Privacy Memory architecture that reframes privacy as user-owned boundary control across social roles. EP-Mem introduces (1) token-level memory driven by user-configurable a privacy policy that stratifies persons and events, combining domain-level default circulation rules with fact-level whitelist/blacklist exceptions; and (2) a pluggable sidecar with a privacy engine that aligns disclosure controls with memory across summary, detail, and boundary granularities, enforced throughout generation, storage, and retrieval. We construct EP-Bench, to our knowledge the first long-term multi-party benchmark with cross-session correlated events for policy-conditioned relational disclosure. Experiments show that EP-Mem achieves 94.0% privacy classification accuracy, improves disclosure-permission judgment from 22% to 68%, and reduces privacy leakage by 75.6%, while maintaining retrieval performance and cross-benchmark generalization. --- ### Beyond Selection: Token Parameterization for Extreme Visual Token Compression Authors: Rui Zhong, Yu Li, Zheyu Yan, Cheng Zhuo Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.35232v2 Visual-token compression is effective for improving the efficiency of vision-language models, but under extreme compression budgets, token pruning can break visual grounding while learned resamplers increase parameter count, attention cost, and training complexity. We revisit compression through a token parameterization lens, separating (i) basis transformation and structured truncation (retained subspace/compressibility) from (ii) coordinate organization (optimization and cross-modal alignment). This view yields two coupled objectives, compressibility and learnability, which we formalize as unified functionals. Guided by these objectives, we design Braco, a lightweight four-step coder that combines transform-basis truncation, input-independent basis-coordinate embeddings, budget-dependent orthogonal re-parameterization, and learned spatial residual tokens from lightweight pooling. Experiments show that Braco forms the favorable empirical accuracy-efficiency frontier under $23\times$--$64\times$ compression and remains competitive at $144\times$, reaching 95.2% accuracy while reducing prefill FLOPs by 84.2%--86.7% relative to the uncompressed upper bound. Against prior methods, Braco matches or improves accuracy while achieving up to approximately 36% end-to-end speedup and using $16.6\times$/$78.8\times$ lower compressor latency/FLOPs. --- ### Token-Disentangled Latent Test-Time Scaling for Vision-Language Reasoning Authors: Hao-Xuan Ma, Yihao Liu, Yutao Sun, Yanting Miao, Mengyu Zhou, YiCheng Xiao, Long Chen, Zhenguo Li, Han-Jia Ye, Xiaoxi Jiang, Guanjun Jiang Published: 2026-09-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.35228v1 Latent test-time scaling improves reasoning by refining hidden states during inference, but existing methods typically apply a single scalar reward to all editable latent tokens. For multimodal large language models, this global update ignores that generated tokens play different roles: some are sensitive to visual evidence, while others correspond to uncertain reasoning decisions. We present Token-Disentangled Latent Test-Time Scaling, an inference-time framework that makes latent refinement token-role-aware. Starting from an initial generated trajectory, we optimize a short hidden-state prefix while routing perception-side visual feedback to image-sensitive tokens and reasoning feedback to high-entropy tokens. Tokens selected by neither route are constrained by an anchor regularizer. Across both perception and reasoning benchmarks on Qwen2.5-VL-7B and InternVL3.5-8B, our method lifts macro accuracy over CoT by +2.57 and +1.51 respectively, and outperforms strong output-space test-time scaling baselines under matched decoded-candidate budgets. Code is available at https://github.com/Qwen-Applications/TD-LTTS. --- ### Generative AI-Based Data Augmentation for Oral Lesion Classification: The PhotoMOCI Dataset and Benchmark Authors: Marco Parola, Mario G. C. A. Cimino, Sabrina Senatore Published: 2026-09-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.35226v1 Early detection of oral cancer via photographic imaging presents a promising avenue for large-scale oral cavity screening. However, the development of robust deep learning models is frequently hampered by the scarcity of high-quality, annotated datasets. To address this limitation, a novel and well-curated resource, the Photographic Multi-purpose Oral Cancer Imaging (PhotoMOCI) dataset, is introduced for developing models across multiple diagnostic tasks in oral oncology. Then, a comprehensive benchmark study was conducted to investigate how various data augmentation strategies influence the performance of image classifiers. Our analysis spans different generative AI frameworks, evaluating the efficacy of traditional methods against advanced generative approaches, including Generative Adversarial Networks (GANs) and Diffusion Models (DMs). Additionally, we propose the Synthetic Image Filter (SIF), a mechanism to select specific samples based on two auxiliary models: Synthetic Proxy Classifier to ensure samples are representative of the target class and Synthetic Image Detector to verify they appear realistic, thereby selecting only the high-utility images that contribute to improving downstream performance. Across the evaluated datasets and classifiers, the best SIF-filtered setup improves accuracy over traditional augmentation in all cases, with gains of +1.73% and +2.35% on PhotoMOCI and +2.38% and +2.08% on KOCD for ResNet50 and ViT, respectively. Our findings reveal that while the direct application of generative data augmentation may yield performance drops, the integration of SIF, considering (i) how synthetic data looks real and (ii) how it reflects the discriminative features of the belonging class, provides a simple yet effective mechanism to filter out synthetic samples that confuse the classifier during training. --- ### SignFLIP: A Unified Model for Sign Language Translation and Generation via Stage-wise Alignment at Scale Authors: Zhaoyi An, Sihan Tan, Youngbae Hwang, Kazuhiro Nakadai, Rei Kawakami Published: 2026-09-28 Categories: cs.CL, cs.CV, cs.MM arXiv: https://arxiv.org/abs/2609.35225v1 Sign language translation and generation share the goal of bidirectional alignment between text and sign representations. However, existing approaches either treat them as isolated tasks or are only verified on limited datasets, limiting effective modeling between modalities. In this paper, we propose SignFLIP, a unified LLM-centered framework for translation and generation. To enable bidirectional mapping between text and sign, SignFLIP adopts a symmetric architecture together with a stage-wise training strategy built on large-scale data. The shared sign--text representation is progressively refined: pre-alignment facilitates subsequent SLT, while the SLT-adapted representation further benefits SLG. Extensive experiments on multiple benchmarks show that SignFLIP shows competitive performance compared with task-specific models on both translation and generation tasks, as well as strong transferability to sign language recognition. --- ### TANGO: Watermarking Masked Diffusion Language Models in Token Pairs Authors: Kasra Arabi, Nir Weinberger, Micah Goldblum, Niv Cohen Published: 2026-09-28 Categories: cs.LG, cs.CL, cs.CR arXiv: https://arxiv.org/abs/2609.35224v1 Masked-diffusion language models fill in masked positions in parallel and in no fixed order. Most practical text watermarks assume left-to-right generation. They key each token to the tokens before it, and in a diffusion model those tokens may still be masked. A fixed green list needs no such context, but it favors the same tokens at every position, so these tokens appear more often in watermarked text. An attacker who compares token frequencies in watermarked and unwatermarked text can recover the list and forge text that the provider's own detector accepts. We present TANGO, a watermark for masked-diffusion language models that keys each new token to a nearby token that is already unmasked. A secret key splits the vocabulary into color classes, and TANGO biases the new token toward a color determined by the key and the nearby token's color. The watermark is therefore embedded in pairs of tokens. Because the favored color changes from position to position, token frequencies stay much closer to those of unwatermarked text than under a fixed green list. Detection needs only the text and the key, and it does not assume any unmasking order. On two masked-diffusion models, TANGO detects nearly all unedited watermarked texts and most edited ones, and frequency attacks that forge the fixed green list fail against it. --- ### Temporal Heterogeneous Graph Pretraining for Relational Deep Learning Authors: Yixin Peng, Er Jin, Diego Collarana, Stefan Decker Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35219v1 Relational deep learning models database rows and foreign-key links as a heterogeneous graph for prediction from record attributes and relational context. These graphs contain two distinct temporal signals: record age changes with the prediction cutoff, while intervals between observed records remain fixed. Prior work often treats time as a single signal or studies temporal representation and pretraining separately. We investigate how explicitly encoding both signals affects temporal pretraining for downstream tasks. Our framework combines Multi-scale Time Encoding, which captures record age using learnable time scales and type-specific projections, with Rotary Time Encoding, which represents signed inter-record intervals through rotary transformations during graph propagation. We pair these encodings with three self-supervised objectives: historical relation recovery, horizon-aware future relation activity prediction, and temporal subgraph contrast. All inputs respect their observation cutoffs. Pretraining proceeds in two stages: subgraph contrast first learns neighborhood representations, followed by refinement through either relation recovery or future activity prediction. We evaluate on five RelBench datasets across 11 classification and regression tasks using heterogeneous GNN and graph Transformer backbones. With both encodings, the best evaluated staged schedules improve over supervised training with the same encodings by 3.02% and 1.06% on the two backbones, respectively, and over controls without pretraining or either encoding by 3.24% and 2.37%. --- ### A Hierarchy of Entropy-Shapley Games for Multivariate Predictive Uncertainty Authors: Niklas Koenen, Claudia Battistin, Jeriek Van den Abeele, Martin Jullum Published: 2026-09-28 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.35217v1 Modern probabilistic machine learning models increasingly produce multivariate outputs with complex dependence structure, from multi-step time-series forecasts to sample path predictions. Understanding which input features drive the predictive uncertainty is important for risk-aware decisions, model diagnostics, and deciding whether the uncertainty should be mitigated or hedged against. This attribution problem requires a choice of how dependencies between output components are treated. Existing approaches reduce the output to a scalar through aggregation or projection before attribution, thereby obscuring whether features affect marginal uncertainty, dependence structure, or both, while component-wise analyses can miss dependence effects entirely. We close this gap by introducing a hierarchy of three entropy-based Shapley games that make this output-side choice explicit for any ordered multivariate outcome, ranging from per-component marginal entropy to fully joint entropy. The hierarchy isolates a cross-component attribution term that captures how each feature shifts the dependence between output components, a quantity invisible to component-wise methods. We establish a chain-rule decomposition of the joint attribution and characterize the cross-component term through conditional total correlation, providing both closed-form and sample-based estimators. Finally, we demonstrate how the framework captures differences in learned joint structure across probabilistic models from distributional regression to a zero-shot time series foundation model. --- ### ASCT: Attentive Search over Counterfactual Trees for Credit Assignment in Agentic Reinforcement Learning Authors: Yang Li, Jinhan Yang, hai liu, Di Wan, Xiyu Chen, Zongsi Xu, Tuo Zhou, Sheng Zhong, Sergey Volkov, Ye Luo, Hao Sun Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.35215v1 Terminal utility evaluates a complete agentic workflow, but learning requires credit for the decisions within it. We introduce Attentive Search over Counterfactual Trees (ASCT), a framework that turns training-time multi-step search into local action credit. At actor-visited states, an auxiliary tree evaluates alternative legal actions from the same recoverable prefix. Its action-value table is centered by the frozen actor's probabilities and supplies credit for PPO on actor-sampled trajectories. This protocol connects counterfactual evaluation to policy learning while deploying the actor alone. Uniform, UCT, and cost-aware AgentUCT instantiate the framework. On HotpotQA agentic retrieval-augmented generation, all three improve mean held-out utility over trajectory-return PPO and workflow-adapted VinePPO. Across three seeds, ASCT-AgentUCT reaches 0.6187 utility versus 0.5939 for VinePPO, with gains in answer F1 and execution cost, and uses 50.3% fewer recorded auxiliary Qwen tokens. Transfer and component-description studies examine the learned policies beyond the training setting. --- ### Uncertainty Quantification in Cardiac Model Personalisation from Ultrafast Ultrasound Authors: Camilla Ferrario, Maelys Venet, Olivier Villemain, Maxime Sermesant Published: 2026-09-28 Categories: q-bio.QM, cs.LG, physics.med-ph arXiv: https://arxiv.org/abs/2609.35214v1 Cardiac model personalisation requires inferring mechanical parameters that are not directly measurable in vivo. Ultrafast ultrasound shear wave elastography (SWE) enables non-invasive tracking of myocardial stiffness dynamics over the cardiac cycle, providing a target for personalisation. However, mapping these observations to subject specific model parameters remains ill-posed, as multiple parameter sets can reproduce the same stiffness dynamics. We formulate SWE-informed personalisation as a statistical inference problem using simulation-based inference (SBI). Using a subject-adapted 0D cardiovascular model and neural posterior estimation, we estimate model-conditional posterior distributions over active stiffness scale k0, contraction rate kATP, and relaxation rate kSR, conditioned on SWE-derived curve features and subject specific context. Among six healthy volunteers, four passed objective prior-support diagnostics and were retained for quantitative posterior analysis. Curve-level RMSE against the observed SWE target decreased from 12.61 $\pm$ 5.55 kPa for the prior predictive median to 1.14 $\pm$ 0.38 kPa for the posterior predictive median, an 89.7 $\pm$ 4.2% reduction. Posterior analysis revealed parameter-specific uncertainty, k0-kATP compensation, weaker constraint of kSR, and the importance of prior-predictive diagnostics for assessing whether each subject is represented within the modelled SWE feature space. These results support SBI for uncertainty aware SWE-based personalisation, while identifying prior support and forward-model adequacy as key diagnostics. --- ### Adversarial Consistency-Guided Representation Learning for Multi-view Clustering Authors: Yuchen Lin, Kunpeng Xu, Ying Fang, Lifei Chen Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.35212v1 Multi-view clustering aims to capture cross-view consistency while exploiting view-specific information. However, shared representations learned to capture cross-view consistency may still retain view-identifying information, potentially compromising the consistency of cross-view clustering structures. To address this issue, we propose ACGRL, an adversarial consistency-guided representation learning framework for multi-view clustering. ACGRL employs a gradient-reversal view discriminator to reduce view identifiability and obtain invariant reference representations. These representations are then frozen to provide fixed references for disentangling view-specific information from cross-view common information in the subsequent learning stage. The fixed reference representations are concatenated with the learned view-specific representations for reconstruction and clustering, with cross-view cluster alignment encouraging consistent clustering assignments. Experiments on four benchmark datasets demonstrate the superior clustering performance of ACGRL compared with representative multi-view clustering methods. --- ### Understanding On-Policy Distillation: A Mechanistic Interpretability Perspective via Sparse Crosscoders Authors: Zichao Yu, Qianshuo Ye, Xu Wang, Difan Zou Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.35210v1 On-policy distillation (OPD) is a widely adopted post-training technique for LLM reasoning. It is commonly believed to transfer knowledge from a stronger teacher, yet what OPD actually distills into the student's internal representations remains unclear. We study this question with sparse crosscoders, which learn one feature dictionary shared by the student before and after OPD and the teacher. Standard crosscoder analyses, however, identify model-specific features but cannot tell how a model's use of its features changes, since all models are encoded into one set of feature activations. We therefore propose the swap readout, which reads each student checkpoint's feature activations on its own, measuring how training changes the student's use of each feature, even for checkpoints unseen by the crosscoder. Across three OPD settings, we find that OPD neither creates features nor passes on the teacher's own, and leaves the firing rates of over 98% of the student's frequently used features within 20%. We further examine the SFT warm-up on the teacher's rollouts that commonly precedes OPD and makes it more effective. Rather than adding features, the warm-up reweights the shared ones in two ways. First, it already raises and lowers many of the features that OPD later raises and lowers, doing part of OPD's work in advance. Second, it changes features that OPD alone would not, notably those for conversation format, reasoning style, and mathematical notation, and these changes persist through OPD. Imposing this reweighting on a directly distilled student's features, without changing its weights, brings its accuracy close to that of the warmed-up student, whereas the same change on shuffled features does not. Together, these findings suggest that OPD reweights existing features rather than acquiring new ones: the student learns from the teacher how to use the features they already share. --- ### From UNI2-h to ConvNeXt-T: Lightweight Nuclei Instance Segmentation via Knowledge Distillation Authors: Wenyan Li Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.35203v1 Nuclei instance segmentation is a core task in digital pathology, yet high-accuracy models rely on large vision transformer (ViT) encoders whose inference speed cannot meet real-time clinical demands. We propose a lightweight scheme that distills the UNI2-h pathology foundation model into a ConvNeXt-Tiny student (Ours-T, 34.7M parameters, 1/20 of the teacher) via output-level knowledge distillation. Ours-T achieves an mPQ of 0.519 on PanNuke (98.8% of the teacher), a zero-shot bPQ of 0.668 on MoNuSeg, and an inference speed of 634.3 img/s, requiring only 0.045 s for full-resolution 1024^2 analysis (21.8x speedup). Experiments further show that multi-scale gated convolution (MALA) yields no gain under ViT encoders, and output-level distillation alone suffices for efficient knowledge transfer. --- ### From Normative Frameworks to Alignment Data: Constructing and Evaluating SFT and Preference Data Authors: Husrev Taha Sencar, Rezart Beka, Danish Naeem, Seda Ozalkan, Majd Hawasly, Ji Lucas, Ala AlFuqaha, Mohamed Abdallah, Recep Senturk Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.35201v1 Aligning language models with a specified normative framework requires translating abstract principles into concrete examples and preference signals from which models can learn. We present an expert-driven methodology for constructing such alignment data and apply it to a normative framework grounded in Islamic ethical, theological, and jurisprudential traditions. Over approximately one year, seven domain experts systematically probed language models to identify alignment deficiencies, curated desired responses, and constructed preference pairs from model outputs and expert judgments. The resulting Arabic-English datasets contain approximately 2.8K supervised fine-tuning (SFT) examples and 5.4K preference pairs spanning a broad range of normative domains. We evaluate the datasets through controlled post-training experiments comparing a Baseline model with models incorporating the curated SFT data alone and both the SFT and preference data. In blind expert evaluation on 150 separately constructed prompts, the model trained with the curated SFT data was preferred over the Baseline in 51.3% of assessor judgments, compared with 14.4% in the opposite direction (p < .001 at the prompt level). Adding the preference data resulted in a smaller difference, with the model trained with both datasets preferred over the SFT model in 28.0% of judgments versus 20.9% in the opposite direction; this difference was not statistically significant at the prompt level (p = .166). Standard Arabic and English benchmarks show no broad degradation in general-purpose capabilities. These results demonstrate how expert-defined normative principles can be systematically operationalized into alignment data and evaluated through controlled model training. --- ### ReCAT: Remember, Count, and Time: Structured Recurrent Memory for Robot Manipulation Authors: Pankhuri Vanjani, Mostafa Hatab, Can Mizrakli, Vaisakh Shaj, Zhuoyue Li, Moritz Reuss, Rudolf Lioutikov Published: 2026-09-28 Categories: cs.RO, cs.AI, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.35200v1 Memory-dependent manipulation requires robots to make decisions using information that is no longer available to their current sensors, such as recalling an earlier visual cue, tracking task progress, counting repeated events, or estimating elapsed time. We present ReCAT, a language-conditioned policy with structured recurrent memory. An instruction-conditioned encoder forms features from the current observation. A recurrent memory integrates the observation stream through Mamba-2 layers and one causal attention layer. A flow-matching Transformer decoder reads the current and the historical representation through separate cross-attention in every block. ReCAT reaches 95.3\% average success on LIBERO and 62.4\% on RMBench, with the best or tied-best result on six of nine tasks. On three real-robot tasks probing spatial recall, event counting, and interval timing, the best ReCAT variant reaches 66.7\% average success, against 8.3\% for the strongest short-history baseline. Controlled comparisons within ReCAT show that the observation encoder and every-block memory conditioning are needed for this performance. They also show that update rules developed for efficient sequence modeling behave differently as robot memory: additive updates have the highest observed success on counting and timing, and delta-rule updates on spatial recall. Project website is at https://intuitive-robots.github.io/ReCAT ---