Knowledge
-
Collaborative Learning and Reasoning over Model Slices
2026-10-09 03:17:29Yuheng Wang, Runde Yang, Zhuohan Bao, Qiran Zhang, Huatao Li, Yiyang Zheng, Lin Wu, Jingru Fan, Shu Yao, Yihan Li, Zhi Li, Chen QianAbstract
Large language models (LLMs) excel across diverse tasks, benefiting from knowledge priors, in-context reasoning, and test-time computation. Despite these strengths, single-model reasoning remains vulnerable to error propagation, as later steps depend on the model's earlier outputs, leaving reasoning blind spots difficult to independently reassess. To address this limitation, we develop \textit{collaborative learning}, encompassing the construction of heterogeneous collaborators and the learning of effective cross-model collaboration. Specifically, \textit{model slicing} replaces a large model with smaller, heterogeneous models under a comparable parameter budget, enabling prior reasoning to be reassessed across distinct model parameterizations. \textit{Pyramid supervision} combines protocol distillation with step-level supervision to refine reasoning and interaction decisions throughout collaboration. We evaluate our approach across four task categories: multi-hop question answering, mathematical reasoning, code generation, and open-ended constrained generation. The trained heterogeneous models achieve comparable or better task performance than the original large model and established multi-agent baselines, including reinforcement-learning-based methods, while also achieving higher inference efficiency than the multi-agent baselines and exhibiting differentiated collaboration patterns across tasks. These findings show that learned collaboration turns model heterogeneity into complementary reasoning capabilities.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12927
PDF
-
A Lightweight Velocity Correction Framework for Improving Instruction Following in Few-Step Text-to-Image Generation
2026-10-09 02:43:49Yating Gao, Zhenlong Zhu, Wen Zihao, Lenan Zhang, Jiexiong Wang, Yang Yang, Guanjun Jiang, Yixiong Zou, Takuya MaekawaAbstract
Few-step text-to-image generators produce high-quality images with few network forward passes, yet struggle with complex prompts involving multiple subjects, interactions, spatial relations, counting, and attribute binding. Can these errors be reduced without changing the generator's parameters or increasing its forward-pass count? We propose a lightweight correction framework for a frozen 6B-parameter generator. The generator predicts velocities that determine how its intermediate image representation is updated; a corrector with fewer than 20M parameters learns residual adjustments at selected sampling steps. Because earlier corrections change the states encountered at later steps, our On-Trajectory Flow-Bridge Supervision derives residual targets at the states actually reached during corrected generation. A single corrector is shared across selected steps and runs in parallel with the generator. Only the corrector is trained, requiring neither reward optimization nor re-distillation. To our knowledge, this is the first framework for lightweight velocity-residual adaptation of fully frozen few-step text-to-image generators that preserves their original sampling schedule and forward-pass count. On our target complex-scene distribution, the framework achieves a +17.60-percentage-point NetWin over frozen Z-Image-Turbo in pairwise instruction-following evaluation, with only 26.9 ms (0.84%) additional end-to-end latency per image and no increase in backbone evaluations.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12899
PDF
-
COFFEQA: Counterfactual-Guided Few-Shot Distillation for LLM Question Answering
2026-10-09 00:41:00Kavinda Kehelella, Sanghamitra DuttaAbstract
Knowledge distillation aims to train a smaller student model to mimic a larger teacher model for resource-constrained environments. However, the process of distillation is typically data-intensive and computationally expensive, particularly for generative tasks. Few-shot distillation offers an alternative by transferring teacher knowledge using limited training samples. However, we observe that few-shot distillation performance is heavily dependent on the choice of these training samples. In fact, we demonstrate that using more uncertain samples near the teacher's decision boundary can significantly deteriorate distillation performance. Motivated by this observation, we propose Counterfactual-guided Few-shot Distillation in Open-book Question Answering (COFFEQA), a novel strategy that aptly uses counterfactuals to provide additional supervision around challenging factual samples. In the generative QA setting, we define a counterfactual as a minimally modified context that induces the teacher to change its response to an alternative answer. We provide geometric insights on how factual-counterfactual pairs can constrain the deviation between the teacher's and student's decision boundaries. We further introduce a new metric, Context Following Difficulty (CFD) that further validates counterfactuals prior to distillation based on agreement with the intended alternative answer and dependence on the modified context. COFFEQA uses CFD-validated factual-counterfactual pairs for few-shot distillation. Experiments across five in-domain and three out-of-domain QA benchmarks using T5 and Qwen2.5 families show that COFFEQA outperforms factual-only distillation baselines under the same few-shot budget. For instance, on TriviaQA, COFFEQA improves over the factual-only baseline by 8.77 average EM points and 9.98 average F1 points under the T5 setting with 32 few-shot pairs.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12837
PDF
-
Tactile Sensing-based Secure and Gentle Grasping of Unknown Objects
Abstract
Reliable robotic manipulation requires maintaining a firm grasp while strictly avoiding both contact slip and structural damage to the object. Satisfying these inherent \textit{secure-and-gentle} grasping constraints is challenging for unknown objects. We present a novel online data-driven predictive control framework for safe dynamic robotic grasping of unknown objects. The proposed system features a pair of compact, single-channel tactile-sensing modules that characterize the gripper-object dynamic interaction in real time, entirely bypassing the need for prior knowledge or offline training. A finite-state-machine-based grasping force probing mechanism is developed to establish the initial grasping force and explicitly identify grasping force constraints. Through online dynamics learning, a Gaussian Process (GP) model captures gripper-object interaction dynamics. The model uncertainties are further used to form robust grasping constraints, which are embedded into a model predictive control scheme. We demonstrate error convergence in a probabilistic sense. The proposed control framework runs at 100~Hz in real time to dynamically manipulate and handle objects rather than static holding. Extensive experimental evaluations across diverse contact surface topologies and various types of objects validate superior performance and efficiency compared to baseline strategies. The proposed method is able to manipulate objects ranging from delicate items, such as raw tofu and marshmallows, to rough objects, such as rocks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12790
PDF
-
EvoBelief: Falsifiable World Models for Self-Improving LLM Agents
2026-10-08 22:46:14Janet Wang, Jiang Guo, Yi Xiang, Jiaqing Yuan, Mingwen Dong, Hideo Kobayashi, Shuaichen Chang, Chandan K. Reddy, Jiarong JiangAbstract
Self-evolving LLM agents typically improve how they act through accumulated experience, reflection, or skills. Yet even procedurally capable agents can fail when their beliefs about environment dynamics are incomplete, incorrect, or outdated. We study a complementary axis of self-evolution: evolving those beliefs during deployment. Such beliefs can support multiple tasks, but incorrect ones can also propagate across them. We introduce EvoBelief, a self-evolving world model that represents environment knowledge as falsifiable hypotheses. Unexpected outcomes and decision-relevant unknowns trigger hypothesis formation. Later observations and controlled experiments provide independent validation before a rule can guide predictions or plans within its supported scope. Counterexamples narrow or retire rules as the environment changes. The language model remains frozen, and each rule remains linked to its evidence and the decisions it changes. Across four benchmarks spanning static dynamics, hidden drift, and scientific discovery, EvoBelief improves task performance and reuses learned rules across tasks. On ALFWorld without actor-side transition memory, it improves success over the matched Frozen World Model by 23.9% and 31.3% with Claude Haiku 4.5 and GPT-5.4, respectively. On MiniGrid, it revises stale rules and recovers after unannounced dynamics changes. In DiscoveryWorld's Reactor Lab, it discovers and validates quantitative laws, completing 83.3% of Normal-tier and 58.3% of Challenge-tier instances, versus 0% for each evaluated baseline. These results demonstrate self-improvement through testable, revisable environment knowledge rather than model-parameter updates.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12789
PDF
-
Faithful and Complete LLM-Assisted Automated Construction of Hierarchical Hardware Knowledge Graphs
2026-10-08 22:12:37Haotian Qiao, Robert P. DickAbstract
Hardware datasheets contain engineering specifications essential for component selection, system integration, and hardware compatibility verification. However, manual inspection is tedious and error-prone because datasheets are lengthy and heterogeneous, with dense information fragmented across tables, figures, and text. Although large language models (LLMs) can automate this process, naïve one-shot processing omits relevant specifications even when the document fits within the model's context window. We present an automated LLM-driven pipeline that transforms hardware datasheets into general-purpose hierarchical hardware knowledge graphs through chunk-based property extraction, hierarchical routing, and iterative refinement. The framework uses LLMs for information extraction and semantic reasoning while preserving source provenance and applying graph updates through validated instructions and deterministic local execution. This staged design reduces per-query context and maintains traceable intermediate representations. The resulting knowledge graphs provide semantically organized, source-grounded hardware properties for downstream retrieval and reasoning. We evaluate the framework on 20 hardware datasheets spanning diverse component types and document lengths. We assess property faithfulness and introduce inter-graph completeness to measure relative semantic coverage without exhaustive ground truth. Our framework achieves 98.8% property faithfulness compared with 76.1% for the one-shot baseline, while fully covering 83.6% of faithful baseline properties; in the reverse direction, the baseline fully covers only 10.5% of ours. These improvements provide more reliable and comprehensive engineering information for downstream hardware design and verification, reducing the risk of overlooked constraints and incorrect decisions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12762
PDF
-
U4RASD at KnowledgeGraphEval-2026: Ontology-Constrained Encoder and LLM Systems for Arabic Relation Extraction
2026-10-08 21:41:29Ahmad Abbas, Hussein Mazeh, Chadi Abou Chakra, Aya Jouni, Mohammad Ballout, Fadi A. ZaraketAbstract
This paper details our submission to KnowledgeGraphEval 2026 Subtask 2 on directed Arabic relation extraction. We explored encoder-based and large language model approaches for predicting one of 40 relations or no_relation from an Arabic sentence and a subject-object pair. Our encoder systems use typed entity markers and separate relation-existence and relation-classification heads, while our LLM systems fine-tune Yehia and Qwen3.5 with QLoRA and ontology-constrained prediction. Our best system, an ensemble of Yehia-RE, Qwen-RE, and TypedPair, achieved a micro-F1 score of 94.06 and ranked 1st in Subtask 2. Yehia-RE (full-vocabulary) achieved 92.98, while the best encoder and Qwen-RE achieved 92.69 and 91.94 respectively. Our results highlight the importance of constrained prediction and complementary modeling approaches for Arabic relation extraction.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12742
PDF
-
KITE: A Knowledge-Injected Temporal Encoder for Time-Series Event Forecasting
2026-10-08 21:35:42Jonas Petersen, Pratik Jawahar, Riccardo Maggioni, Gian-Alessandro Lombardi, Matteo Cozzi, Marek Gajewski, Federico Martelli, Camilla Mazzoleni, Philipp Christian PetersenAbstract
Forecasting rare, high-cost events in multivariate time-series systems is hard because their precursors are seldom observed, so a purely data-driven model has little to learn from. The engineers who run these systems, however, hold knowledge that precedes such events, information absent from the scarce labels. We call this a semantic prior and inject it into a self-supervised event forecaster through two switchable mechanisms, a covariate mechanism and an attention-bias mechanism, each the identity at its null, so one architecture serves every prior type with no per-dataset architecture search. A prior helps when it supplies information the observed window does not already contain: in a short context an informative prior beats no prior in-domain (pooled h-AUROC over the no-prior baseline: a periodicity clock +14%, p<0.001; a conservation residual +47% at a 0.25 label fraction), and a working prior transfers out-of-domain (+53%), substituting for feature transfer, while priors the window already recovers stay inert and the attention-bias mechanism shows no measurable gain. KITE is a switchable, attributable route to contextual conditioning, and a substrate for scaling toward cross-domain, pretrained time-series foundation world models that carry their governing structure explicitly.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12738
PDF
-
Activating the Annotator: Operationalising Lived Experience in Data Annotation
2026-10-08 20:31:56Sheza Munir, Ramaravind Kommiya Mothilal, Syed Ishtiaque AhmedAbstract
Perspectivist NLP has established that annotator positionality shapes labels for subjective tasks, but the field still lacks methods for operationalising it. Positionality is typically operationalised as a demographic form completed at recruitment: a static attribute recorded once and treated as a covariate, on the assumption that the knowledge it indexes is active when the annotator reads an item. This paper treats that assumption as a design problem and asks whether lived experience can instead be primed by the task itself. We introduce AXLE, a pre-annotation stage in which a scenario vignette anchors a short, axis-guided conversation that is synthesised into a first-person micro-narrative and returned to the annotator before labelling begins. In a 2x2 between-subjects study (104 annotators, 30 items, 1,040 annotations) on ambiguous hate speech spanning immigration, religion and gender, elicitation was associated with higher identity resonance and higher sensitivity to harm. It also increased label entropy and lowered inter-annotator agreement: the stage appears to surface plurality rather than manufacture consensus, which distinguishes it from annotator training, whose success criterion is convergence.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12701
PDF
-
Spa-Bench: A Real-Robot Benchmark for Spatial Grounding in Vision-Language-Action Models
2026-10-08 20:22:35Justin Tien-Smith, Matthew Pidden, Rolandos Alexandros PotamiasAbstract
Vision-language-action models (VLAs) offer a promising route toward general-purpose robotics by transferring semantic knowledge from large-scale vision-language pretraining into language-conditioned control. However, most VLA evaluations emphasize aggregate manipulation success under changes in objects, instructions, or environments, leaving unclear how reliably these policies ground spatial descriptions in the physical scenes where they act. We therefore ask a sharper question: how reliably can current VLAs translate spatial descriptions into appropriate real-robot behavior? To investigate this, we introduce Spa-Bench, a real-robot benchmark that evaluates spatial grounding across six tabletop tasks covering Object State, Relative Size, Referential Description, Relational Placement, Ordinal Position, and Counting. Our primary evaluation compares in-distribution (ID) and out-of-distribution (OOD) spatial configurations while retaining training support for the relevant objects, concepts, and actions, reducing object and motor unfamiliarity as alternative explanations for failure. Four diagnostic conditions further probe physical execution, sensitivity to wording, object-task associations, and recognition that no action is required. We release 4,092 episodes collected on the SO-101 robot arm: 1,200 human-teleoperated demonstrations and 2,892 real-robot evaluation rollouts from five state-of-the-art VLA policies. Across the three fully evaluated policies, success falls from 67.5-73.3% on ID spatial configurations to 40.7-52.0% on OOD spatial configurations, while matched direct instructions improve success by 28.3-40.0 percentage points. These results show that the evaluated VLAs can often execute the required manipulations but lack robust spatial grounding. Project materials, including videos, code, and benchmark data, are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12696
PDF
-
Incremental Retrieval-Augmented Generation for Human-in-the-Loop Decision Support
Abstract
We introduce iRAG, a human-in-the-loop decision-support system combining Retrieval-Augmented Generation (RAG) with Hybrid Decision-Making. Unlike conventional RAG systems, relying on static Knowledge Bases (KBs), our proposal addresses settings where the KB is initially empty and is progressively populated through human decisions. The system stores previous decisions as retrievable precedents and moves from silent observation to evidence-grounded contestation and, finally, to autonomous handling with abstention. Against an independent baseline, iRAG matches final-decision error under 10% drift without retrieval decay and reduces it under 40% drift across all configurations. Retrieval decay improves performance on changed knowledge while increasing error on stable knowledge, evidencing a stability--plasticity trade-off rather than a general advantage. Human oversight is an architectural provision of the workflow; its usability with real human users remains to be evaluated.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12666
PDF
-
GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence Reports
2026-10-08 18:35:13Veronika Solopova, Ivan Havlytskyi, Viktoria Skorik, Ostap VykhopenAbstract
Geopolitical and strategic intelligence reports constitute a rich source of expert knowledge and have increasing value in the current political climate, but their heterogeneity and analytical complexity pose challenges for retrieval-augmented question answering. We introduce a benchmark for grounded question answering over 415 open-source geopolitical and strategic intelligence reports, comprising 5,525 document-grounded question-answer pairs spanning factual, inferential, comparative, and risk-oriented questions. We propose a document-aware generation pipeline that profiles reports before generating questions and validate dataset quality through expert annotation and an LLM-as-judge protocol. Using the benchmark, we evaluate 15 retrieval systems and six open-weight LLMs for RAG. We find that models perform similarly on document-level retrieval, while at the harder chunk level retrieval quality is not driven by embedding-model scale: bge-m3, a mid-size open encoder matches an order of magnitude larger ones and outperforms commercial embedding APIs. In contrast, answer generation remains challenging even given the gold evidence, particularly for inferential and risk-oriented questions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12617
PDF
-
Knowing Is Not Telling: Identity Ossification in LLM Representations of Trans-border Communities
2026-10-08 18:31:13Minglu Sun, Lan Chen, Hang JiangAbstract
Ethnic and cultural identities rarely align with political borders, a tension made sharper by migration and long-standing cross-border ties. Communities such as the Sami, the Rohingya, and the Kurds have long lived across more than one state. This tension is especially clear in mainland Southeast Asia, where many communities share history, language, and culture across state lines. We study nine such transborder ethnic communities along China's southwestern border with mainland Southeast Asia, looking at how large language models (LLMs) represent them. We define and study identity ossification: a pattern where LLMs turn a fluid, cross-border ethnic identity into a single, fixed nation-state identity. We introduce the Transborder Identity Ossification Diagnostic Framework (TIODF) that compares direct factual questions with open-ended narrative questions to test whether ossification comes from missing knowledge or from how models frame their answers. We test GPT-5.1 and DeepSeek-V3.2 in English and Chinese. Ossification varies substantially across models and query languages. It appears in only 10.1% of GPT-5.1's English answers, but in 62.6% of DeepSeek-V3.2's Chinese answers. Models often still tell a single-nation story even when they answer the factual question correctly. This decoupling suggests that ossification reflects narrative framing rather than simple knowledge absence, which matters for how we test cultural fairness in multilingual AI systems.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12610
PDF
-
SpireBench: Evaluating Long-Horizon Strategic Play of Agentic Models with Slay the Spire
2026-10-08 18:18:19Scott C. Lowe, Sander W. KeeminkAbstract
Agentic large language models now tackle complex tasks, yet strategic reasoning over long horizons is under-studied: most agentic benchmarks are short and episodic, and performance on long tasks is hard to dissect into stored knowledge, reasoning and luck. We introduce SpireBench, in which models play the complete commercial deck-builder Slay the Spire (StS), where early decisions decide long-term success. Across 20 large language models (LLMs) at the game's highest difficulty, frontier models beat heuristic baselines and match the average human player but remain far behind experts. This gap widens as the run lengthens: the strongest models win the first act as often as experts, then fall behind. What separates strong models from weak is how well they play each fight; all of them misjudge risk over a run, preserving health rather than spending it to grow stronger. A comprehensive, self-consistent renaming of every game term shows that part of their performance rests on prior knowledge of the game, and models do not improve from the memories they write across runs, which are often wrong and overly cautious. Our harness is the first to give a model the complete original game, and we compare it with existing harnesses. Overall, SpireBench offers an unsaturated long-horizon benchmark with tools to locate where and why agents fail.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12596
PDF
-
Scaling Multi-Teacher Distillation for Digital Pathology
2026-10-08 18:02:09Sofi\`ene Boutaj, Pierre Marza, Varun Belagali, Dimitris Samaras, Maria Vakalopoulou, Stergios ChristodoulidisAbstract
Multi-teacher distillation allows transferring knowledge from multiple "teacher" networks to a single "student" network. This method is promising in fields such as digital pathology, where many powerful foundation models were proposed recently. However, as we show in this paper, the standard multi-teacher distillation approach reacts poorly to an increase in the number of teachers used to train a student encoder. We demonstrate that learning teacher-specific representations is a key to scaling in MTD. Importantly, different design choices, such as learnable teacher tokens, a tailored attention scheme, additional mixture-of-experts layers and a contrastive loss, are proposed to better learn such teacher-specific representations. We show that our method better scales with respect to the number of considered teachers, allowing us to train compact student encoders, up to 10x more efficient than larger teacher foundation models, while matching their performance and even outperforming them on a large set of tile-level and slide-level tasks. We conduct a thorough empirical validation, evaluating more than 10 foundation models on 39 tasks spanning tile and slide levels. As a result, we release a new collection of strong and efficient foundation models, named OMNI, trained from the knowledge of 10 state-of-the-art foundation models. We finally conduct an analysis of the learned teacher-specific representations, highlighting their complementarity and explaining why they can be easily aggregated at downstream time.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12574
PDF
-
Caught in the Act: Probes Effectively Detect Sabotage and Catch Unverbalized Deception
2026-10-08 17:58:13Oskar J. Hollinsworth, Alex F. Spies, Tigist Diriba, Adam Gleave, Chris CundyAbstract
Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception detection via probes can be scaled up to frontier monitoring settings by collecting the largest deception dataset to date for training probes and introducing a novel probe architecture which can aggregate information across many layers and tokens. Our probes achieve 98.8% AUC in SHADE-Arena, surpassing an Opus 5.5 text-monitoring baseline, and show improved efficacy as the underlying model is scaled up. To push our probes to their limit, we test them on several cases where deception cannot be determined from the context alone. In these cases, which we refer to as introspective deception, the ground truth can only be determined through careful elicitation or thorough knowledge of a model's training data. In one such evaluation, we show that probes can distinguish transcripts containing a model's true hidden goal from other goals with an AUC of up to 99.7%. Our probes also readily detect deception on prominent open-weight models which lie about politically sensitive topics, and about their beliefs when put under pressure. We release our training dataset, dubbed FIBS, to help drive frontier deployment of effective probes, and encourage the community to expand upon it with further examples of deception and sabotage.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12445
PDF
-
Humanity's Sixth Sense: Benchmarking Intuitive Visual Reasoning in Multimodal Models
2026-10-08 17:39:44Xingang Guo, Jing Gu, Brian Jang, Renxiong Wang, Utkarsh Tyagi, Daniel Quigley, Steven Li, David Yan, Daniel Yue Zhang, Darvin Yi, Forrest Huang, HiJae Kim, Tianyi Zhang, Jared Lichtarge, Jihua Huang, Le Xue, Manan Tomar, Qiuyi Richard Zhang, Ruofei Yu, Seth Neel, Yaning Hu, Marcella Valentine, Xinzhe Jiang, Daniel Evans, Chenguang Wang, Dustin Tran, Tong Zhao, Yinfei Yang, Yunzhong HeAbstract
Humans perceive far more in a scene than what is explicitly depicted: a single glance captures past causes and future trajectories; a quick peek determines if a vehicle can fit between two parked cars; a few seconds of video reveals who holds authority in a room; and a fleeting clip highlights subtle abstract patterns like unwritten rules or hidden labels. This capacity reflects a form of humanity's sixth sense: an intuitive reasoning mechanism that recovers implicit information beyond raw sensory perception. Crucially, this rapid, zero-shot visual intuition underpins everyday navigation and social interaction, making it a vital capability for Multimodal Large Language Models (MLLMs) deployed alongside people. Existing visual benchmarks, however, target either deliberate expert-level analysis in academic and mathematical domains or low-level perception, leaving the intuitive reasoning that people perform largely untested. To bridge this gap, we introduce Humanity's Sixth Sense (HSS), a benchmark for intuitive visual reasoning. HSS spans diverse image and video inputs, organizes items under a structured taxonomy, and pairs each with human-written prompts probing the implicit temporal, spatial, social, and abstract structure that people infer at a glance. Frontier MLLMs fall short of human performance: participants reach 93.1% accuracy, while the strongest model, GPT-6-astra, reaches only 53.6% even at maximum reasoning effort. Despite excelling in many complex tasks that require advanced perception and knowledge, current models still struggle significantly on these visual tasks that are intuitive for humans. We further explore agentic setup that apply dynamic visual manipulation to HSS, which narrows but does not close the gap. HSS establishes intuitive visual reasoning as a measurable axis and directs attention to a capability that scaling on current benchmarks has so far left behind.
Abstract (translated)
URL
https://arxiv.org/abs/2610.08966
PDF
-
ARC: A Reasoning Recipe for Robot Foundation Models
2026-10-08 17:38:04Gokul Puthumanaillam, Tao Sun, Elie Aljalbout, Moritz Reuss, Zhaoshuo Li, Fabio Ramos, Ankit Goyal, Jenai Xuning YangAbstract
The prevailing approach to improving robot foundation models (RFMs) relies on larger models, more robot demonstrations, and costly training at scale. We show that there exists an effective and efficient complementary approach: the right reasoning recipe can substantially improve the zero-shot task performance of existing state-of-the-art RFMs. We refer to this recipe as ARC. It consists of three key ingredients: a reasoning trace, a scalable automatic labeling pipeline, and a strategy for adapting pretrained RFMs to use these traces for control. First, we find that effective reasoning traces should be grounded in the robot's next action and explain its causal structure: why the action is appropriate and what effect it should produce. Second, we show that these traces can be generated automatically from existing demonstrations, enabling us to construct ARC-Trace-DROID from DROID without collecting new robot data. Third, we show how state-of-the-art VLAs such as $\pi_{0.5}$ and WAMs such as Cosmos3-Nano-Policy can learn to use these traces for control, with fine-tuning and inference tailored to each model's architecture and capabilities. Using ARC, we obtain gains in zero-shot RFM performance that, to our knowledge, are unprecedented without additional robot demonstrations or foundation-scale training. The adapted models establish a new state of the art on RoboLab-120 and MolmoSpaces, with gains of up to 50 percentage points on RoboLab-Reasoning-50. On real robots, ARC improves $\pi_{0.5}$'s task success by 82.2 percentage points. Project website: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2610.12386
PDF
-
OnTrack: Real-Time Monitoring and Intervention in LLM Agent Trajectories via Streaming Structure-Aware Optimal Transport
2026-10-08 17:32:08Babak Barazandeh, Connor Swanson, Chinmay Kulkarni, Nikhil MungelAbstract
Agents are deployed in applications from trip planners and stock trading to IT incident triage. In most cases, LLM agents work autonomously with minimal rule-based safeguarding, leading to cost and safety issues from irreversible actions. Recent works resolve this either by using a safeguard agent to monitor behavior or evaluating logs post-hoc. The first adds cost and latency to every step; the second delivers its verdict after the run, when tokens are burned and damage is done. To overcome this, we propose OnTrack, a streaming monitoring mechanism that compares an agent's steps and dependencies against recorded successful runs to alert users or block the agent in about a millisecond per step. We study this problem in three regimes of decreasing access: full reference access (historical runs and tool schemas), intermediate access (only tool schemas), and no prior knowledge (only step logs as generated). Expectation of OnTrack's monitoring capabilities reduces as data access drops, ranging from plan violation detection to identifying loops, stalls, and repeated tool calls. Finally, we evaluate OnTrack using SWE-bench trajectories. Based on the first 8 steps, our method ranks failing trajectories below succeeding ones better than content similarity approaches (+0.057 AUROC). With an abort policy, we save about 18% of compute that would be burned on failing runs, where 83% of interrupted runs were actually heading to failure (5 out of 6 aborts were correct).
Abstract (translated)
URL
https://arxiv.org/abs/2610.12375
PDF
-
Accurate but Not Humble: Evaluating Epistemic Humility in LLM Agents under Knowledge Conflict
2026-10-08 17:25:06Kaiser Sun, Bernal Jimenez Gutierrez, Hongjun Liu, Jingyu Zhang, Jie Gao, Mark Dredze, Daniel KhashabiAbstract
When retrieved evidence contradicts an agent's prior beliefs, does it revise its answer, acknowledge uncertainty, or persist with an incorrect conclusion? Existing evaluations of agentic systems focus primarily on task success, offering limited insight into how agents handle such conflicts. We propose to evaluate agents on epistemic humility (EH): the agent's willingness to recognize, act on, and communicate uncertainty during task execution. We operationalize EH through three trajectory-level behavioral dimensions: Identify, Solve, and Escalate (ISE). Through knowledge conflict, situations where the backbone language model's parametric knowledge contradicts the evidence it encounters, or where two contextual sources disagree, we evaluate two conflict settings: (1) controlled conflict and (2) naturally occurring conflict during multi-step agentic execution, each paired with matched no-conflict controls. Evaluating four agents, we find that higher task accuracy does not necessarily correspond to greater epistemic humility: some high-accuracy configurations recognize conflicts during execution but do not communicate unresolved uncertainty in their incorrect final answers. Trajectory-level analysis further reveals that agents frequently detect conflicts in early steps of execution but fail to maintain or resolve them in later steps. Finally, we show that model-level interventions can improve EH, but often at the cost of task accuracy, suggesting that epistemic humility emerges from the interaction among the backbone model, agent harness, and evaluation environment.
Abstract (translated)
URL
https://arxiv.org/abs/2610.12360
PDF
