Every paper checked — nothing hidden

All papers

Every paper checked for brief #4 of 12–19 Oct 2026 — nothing is hidden. Brief papers first.

All17,313 In brief24 More (on-topic)3,333 Not relevant13,754 Not assessed87 Older352 Could not confirm93 Blogs with no paper266 All blogs284

Paper	Topic	Relevance
Mitigating Capability Degradation in LLM Self-Improvement via Representation Geometry Regularization Introduces representation geometry regularization to prevent capability degradation during on-policy self-distillation in language models. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Proposes a training regularization technique, not an LLM evaluation framework or benchmark study.
Localized Generative Design of Ship Hulls with Diffusion Models and High-Fidelity Assessment Applies a reward-directed diffusion model to optimize localized ship hull geometry evaluated by automated fluid dynamics simulations. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on ship hull design using diffusion models, not multi-step LLM agents or tool use.
Localized Generative Design of Ship Hulls with Diffusion Models and High-Fidelity Assessment Applies a reward-directed diffusion model to optimize localized ship hull geometry evaluated by automated fluid dynamics simulations. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Covers CFD evaluation for ship hulls, not LLM evaluation benchmarks or methodologies.
Localized Generative Design of Ship Hulls with Diffusion Models and High-Fidelity Assessment Applies a reward-directed diffusion model to optimize localized ship hull geometry evaluated by automated fluid dynamics simulations. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Concerns fluid dynamics and physical constraints in engineering, not AI safety or deployment guardrails.
Collaborative Theory of Mind: Learning to Self-Revise by Anticipating User Responses Trains language models to anticipate user replies and self-revise drafts before responding in multi-turn interactions. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Proposes a training framework for collaborative revision rather than a benchmark.
Collaborative Theory of Mind: Learning to Self-Revise by Anticipating User Responses Trains language models to anticipate user replies and self-revise drafts before responding in multi-turn interactions. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on multi-turn user collaboration performance rather than safety guardrails or jailbreaks.
One-Step Text Generation by Seq-Drifting Proposes a sequence-drifting framework to generate full text sequences in a single forward pass from Gaussian noise. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Investigates one-step text generation architectures rather than deployment safety or robustness.
One-Step Text Generation by Seq-Drifting Proposes a sequence-drifting framework to generate full text sequences in a single forward pass from Gaussian noise. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Introduces a text generation framework rather than an evaluation benchmark or judge system.
One-Step Text Generation by Seq-Drifting Proposes a sequence-drifting framework to generate full text sequences in a single forward pass from Gaussian noise. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Proposes a non-autoregressive text generation diffusion method, not autonomous agents.
Mind the Spatial Gap: A Benchmark for Geospatial Reasoning and Constrained Spatial Editing Presents a benchmark using real urban data to evaluate multimodal language models on spatial reasoning and constrained editing. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Evaluates geospatial reasoning accuracy rather than safety, guardrails, or reliability risks.
Mind the Spatial Gap: A Benchmark for Geospatial Reasoning and Constrained Spatial Editing Presents a benchmark using real urban data to evaluate multimodal language models on spatial reasoning and constrained editing. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Presents a benchmark for spatial reasoning and editing, not multi-step tool-using agents.
OGPO: Offline Goal-conditioned Policy Optimization for Recoverable Vision-Language-Action Models Introduces an offline goal-conditioned policy optimization method to help vision-language-action models recover from execution deviations. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on robotic trajectory recovery rather than AI safety guardrails or jailbreaks.
OGPO: Offline Goal-conditioned Policy Optimization for Recoverable Vision-Language-Action Models Introduces an offline goal-conditioned policy optimization method to help vision-language-action models recover from execution deviations. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents policy optimization for robotics rather than LLM evaluation or benchmark design.
DeFiRank: Federated Pairwise Ranking for Governance in Decentralized Financial Networks Proposes a federated pairwise ranking framework for institutions to evaluate decentralized finance participants without sharing transaction data. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Focuses on a federated ranking model for financial transactions rather than LLM benchmarks.
DeFiRank: Federated Pairwise Ranking for Governance in Decentralized Financial Networks Proposes a federated pairwise ranking framework for institutions to evaluate decentralized finance participants without sharing transaction data. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Deals with DeFi governance ranking rather than LLM safety, red-teaming, or guardrails.
DeFiRank: Federated Pairwise Ranking for Governance in Decentralized Financial Networks Proposes a federated pairwise ranking framework for institutions to evaluate decentralized finance participants without sharing transaction data. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Proposes federated pairwise ranking for blockchain governance, not LLM agents.
AI-Mediation in Retail Markets: Revenue Effects and Assortment Optimization Studies the impact of AI shopping assistant mediation on retail product assortment planning and business revenue. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on retail revenue optimization rather than deployment safety, guardrails, or hallucination.
AI-Mediation in Retail Markets: Revenue Effects and Assortment Optimization Studies the impact of AI shopping assistant mediation on retail product assortment planning and business revenue. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Analyzes retail assortment optimization rather than benchmarking or evaluating LLMs.
AI-Mediation in Retail Markets: Revenue Effects and Assortment Optimization Studies the impact of AI shopping assistant mediation on retail product assortment planning and business revenue. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Economic study of retail assortment planning under AI recommenders, not agent architectures.
Teaching neural decoders with language Uses natural language class descriptions as supervision to train neural decoders for brain-computer interfaces across multiple modalities. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Proposes a supervised decoding framework for neural signals, not LLM benchmarking.
Teaching neural decoders with language Uses natural language class descriptions as supervision to train neural decoders for brain-computer interfaces across multiple modalities. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on BCI decoding performance rather than deployment safety or model alignment.
Teaching neural decoders with language Uses natural language class descriptions as supervision to train neural decoders for brain-computer interfaces across multiple modalities. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Deals with brain-computer interface neural decoders rather than LLM agents.
Collaborative Learning and Reasoning over Model Slices Develops a collaborative learning framework that splits large language models into smaller heterogeneous slices to mitigate error propagation. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Studies multi-model reasoning collaboration rather than deployment safety or jailbreak prevention.
Collaborative Learning and Reasoning over Model Slices Develops a collaborative learning framework that splits large language models into smaller heterogeneous slices to mitigate error propagation. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Focuses on a multi-model reasoning framework rather than benchmarking or evaluation metrics.
Learning More by Training Less: Uncertainty-Guided Progressive Contraction for Efficient Reasoning Post-Training Presents an uncertainty-guided framework to compress reasoning supervision trajectories and reduce reasoning post-training costs. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on reasoning post-training data selection, not autonomous agent execution.
Learning More by Training Less: Uncertainty-Guided Progressive Contraction for Efficient Reasoning Post-Training Presents an uncertainty-guided framework to compress reasoning supervision trajectories and reduce reasoning post-training costs. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on training efficiency and reasoning quality rather than safety or reliability.
Learning More by Training Less: Uncertainty-Guided Progressive Contraction for Efficient Reasoning Post-Training Presents an uncertainty-guided framework to compress reasoning supervision trajectories and reduce reasoning post-training costs. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Introduces a post-training compression framework rather than a testing or benchmark methodology.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Studies layer selection for speculative decoding, not code generation or program repair.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Deals with inference acceleration rather than document retrieval or grounded generation.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on inference efficiency rather than safety, reliability, or red-teaming.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Studies text LLM hidden states and speculative decoding without multimodal elements.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on speculative decoding acceleration, not autonomous agent planning or actions.
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding Proposes matrix-entropy and kernel alignment criteria to select optimal layer subsets of hidden states for speculative decoding. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Analyzes internal layer information for inference speedup rather than model benchmarking.
MathVis: Can AI Reason through Drawing? A Benchmark for Mathematical Figure Generation Presents a benchmark to evaluate language models on generating mathematical figures and diagram-assisted reasoning across frontier problems. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Addresses mathematical reasoning evaluation, not deployment safety, guardrails, or robustness.
MathVis: Can AI Reason through Drawing? A Benchmark for Mathematical Figure Generation Presents a benchmark to evaluate language models on generating mathematical figures and diagram-assisted reasoning across frontier problems. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Evaluates model reasoning via drawing rather than autonomous multi-step agent systems.
JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents Introduces a multimodal embodied agent policy learning framework using trajectory construction, distillation, and token-level evidence routing. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents policy learning and distillation rather than new benchmarks or evaluation methodologies.
JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents Introduces a multimodal embodied agent policy learning framework using trajectory construction, distillation, and token-level evidence routing. Preprint	Computer vision for industrial applications via arxiv · 12 Oct, 16:28	Not relevant Focuses on robotics manipulation policy learning rather than industrial visual inspection.
JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents Introduces a multimodal embodied agent policy learning framework using trajectory construction, distillation, and token-level evidence routing. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Uses distillation for policy learning rather than small model compression or efficient serving.
JoyInside-VLM-Agent: Product-Conditioned Multimodal Policy Learning for Embodied Agents Introduces a multimodal embodied agent policy learning framework using trajectory construction, distillation, and token-level evidence routing. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Addresses policy task performance rather than safety guardrails, jailbreaks, or red-teaming.
MotionCoach: Constructing Multimodal Guidance for Video Professional Motion Editing Proposes a multimodal agent providing grounded prompts and video clips to guide video editing models in professional sports motions. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Uses an agent workflow as a tool for video editing rather than studying autonomous agents.
MotionCoach: Constructing Multimodal Guidance for Video Professional Motion Editing Proposes a multimodal agent providing grounded prompts and video clips to guide video editing models in professional sports motions. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on motion editing quality rather than safety, robustness, or alignment risks.
MotionCoach: Constructing Multimodal Guidance for Video Professional Motion Editing Proposes a multimodal agent providing grounded prompts and video clips to guide video editing models in professional sports motions. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Introduces VPME for evaluation, but core contribution is a video motion editing method.
MotionCoach: Constructing Multimodal Guidance for Video Professional Motion Editing Proposes a multimodal agent providing grounded prompts and video clips to guide video editing models in professional sports motions. Preprint	Computer vision for industrial applications via arxiv · 12 Oct, 16:28	Not relevant Focuses on sports and performance video editing rather than industrial inspection or manufacturing.
SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents Introduces a multi-turn conversational benchmark to evaluate sales agents against resistant, psychologically grounded buyer personas. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Benchmark paper with no focus on model quantization, distillation, or efficient serving.
SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents Introduces a multi-turn conversational benchmark to evaluate sales agents against resistant, psychologically grounded buyer personas. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Evaluates sales dialogue agents, not models that write, review, or fix code.
SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents Introduces a multi-turn conversational benchmark to evaluate sales agents against resistant, psychologically grounded buyer personas. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Uses RAG internally for dialogue grounding but does not contribute to RAG methods.
SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents Introduces a multi-turn conversational benchmark to evaluate sales agents against resistant, psychologically grounded buyer personas. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Evaluates text-based multi-turn conversational agents without multimodal modalities.
SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents Introduces a multi-turn conversational benchmark to evaluate sales agents against resistant, psychologically grounded buyer personas. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Evaluates commercial sales performance and persistence rather than safety or guardrails.
Beyond Semantic IDs: Finite-Beam Reachability and Continuous-Query Generation in Generative Recommendation Proposes continuous-query generation paradigms to eliminate reachability constraints caused by finite-beam decoding over discrete semantic IDs in recommendation. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on recommender retrieval reachability, not AI safety or deployment guardrails.
Beyond Semantic IDs: Finite-Beam Reachability and Continuous-Query Generation in Generative Recommendation Proposes continuous-query generation paradigms to eliminate reachability constraints caused by finite-beam decoding over discrete semantic IDs in recommendation. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents a continuous-query recommendation architecture rather than LLM evaluation benchmarks.
Beyond Semantic IDs: Finite-Beam Reachability and Continuous-Query Generation in Generative Recommendation Proposes continuous-query generation paradigms to eliminate reachability constraints caused by finite-beam decoding over discrete semantic IDs in recommendation. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Concerns recommendation system retrieval architecture, not LLM agents taking environment actions.
The Evidence Anchoring Boundary: Where Visual Interventions Matter and When They Help in Multimodal LLMs Analyzes layer-wise visual interventions in multimodal language models to identify where visual edits effectively reduce hallucinations. Preprint	Computer vision for industrial applications via arxiv · 12 Oct, 16:28	Not relevant Focuses on general multimodal LLM hallucination rather than industrial computer vision inspection.
The Evidence Anchoring Boundary: Where Visual Interventions Matter and When They Help in Multimodal LLMs Analyzes layer-wise visual interventions in multimodal language models to identify where visual edits effectively reduce hallucinations. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Investigates intervention mechanics and hallucination reduction rather than proposing a benchmark.
The Evidence Anchoring Boundary: Where Visual Interventions Matter and When They Help in Multimodal LLMs Analyzes layer-wise visual interventions in multimodal language models to identify where visual edits effectively reduce hallucinations. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Investigates internal representation interventions in vision-language decoders, not agent planning.
Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction Introduces a framework and multimodal dataset of human desktop demonstrations to induce reusable skills for computer-use agents. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Primary focus is skill induction for agents, though it includes an evaluation dataset.
Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction Introduces a framework and multimodal dataset of human desktop demonstrations to induce reusable skills for computer-use agents. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Addresses GUI task automation rather than deployment safety, jailbreaks, or guardrails.
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training Introduces an information-theoretic framework to reweight supervised fine-tuning data based on conditional information gain. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Pertains strictly to language model post-training without multimodal components.
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training Introduces an information-theoretic framework to reweight supervised fine-tuning data based on conditional information gain. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Addresses general supervised fine-tuning techniques, not agent planning or tool use.
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training Introduces an information-theoretic framework to reweight supervised fine-tuning data based on conditional information gain. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Focuses on SFT token weighting rather than retrieval or grounded knowledge bases.
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training Introduces an information-theoretic framework to reweight supervised fine-tuning data based on conditional information gain. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Addresses generalization during SFT rather than safety, hallucination mitigation, or robustness.

13,381–13,440 of 13,754← Previous Next →