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
Who May Release an Agent's Promise? Standing over Commitments in Multi-Party Chat Studies whether conversational agents correctly enforce authorization rules when users attempt to cancel prior agent commitments to others. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Focuses on multi-party conversation, not code generation.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Analyzes linguistic interpretation of logical terms, not safety risks.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Studies reasoning chain effects on semantics, not autonomous agents.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Does not explore model compression or efficient serving.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Examines text reasoning models, not multimodal inputs.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Does not involve external retrieval or document databases.
Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally Examines how extended reasoning steps lead language models to interpret everyday logical terms strictly literally instead of inferring pragmatic meaning. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Focuses on natural language pragmatics, not software code.
Rubric Judgments Lack Uncertainty: Conformal Weighting for LLM Evaluation Proposes conformal calibration to weight criterion-level uncertainty when using rubric-based language model judges for reinforcement learning. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Addresses rubric aggregation for LLMs, not multi-step agent actions.
AutoOnboarder: Autonomously Onboarding LLM Agents Presents an autonomous supervisor agent that iteratively tunes an agent's code, harness, and prompt on few examples. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Presents an agent optimization framework, not an evaluation standard.
AutoOnboarder: Autonomously Onboarding LLM Agents Presents an autonomous supervisor agent that iteratively tunes an agent's code, harness, and prompt on few examples. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on performance and cost optimization rather than safety.
RecSoup: Efficient Model Merging Approach for Controllable Multi-Objective Recommendation Introduces a two-stage model merging method to enable dynamic multi-objective recommendation trade-offs at inference time without retraining. Peer-reviewed	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Addresses recommendation model merging, not retrieval-augmented text generation.
RecSoup: Efficient Model Merging Approach for Controllable Multi-Objective Recommendation Introduces a two-stage model merging method to enable dynamic multi-objective recommendation trade-offs at inference time without retraining. Peer-reviewed	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Does not focus on deployment safety, jailbreaks, or hallucinations.
RecSoup: Efficient Model Merging Approach for Controllable Multi-Objective Recommendation Introduces a two-stage model merging method to enable dynamic multi-objective recommendation trade-offs at inference time without retraining. Peer-reviewed	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes a model merging technique rather than an evaluation benchmark.
RecSoup: Efficient Model Merging Approach for Controllable Multi-Objective Recommendation Introduces a two-stage model merging method to enable dynamic multi-objective recommendation trade-offs at inference time without retraining. Peer-reviewed	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on recommender system parameters, not autonomous agents.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Investigates internal representations of sparse autoencoders, not agents.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Focuses on feature sensitivity analysis, not LLM benchmarking.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Does not address external document retrieval or grounding.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Studies feature interpretability, not model compression or serving efficiency.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Does not address coding tasks or software engineering.
Representation Redistribution in Sparse Autoencoders Analyzes how prompt variations redistribute internal feature activations in sparse autoencoders and traces their impact on downstream model decisions. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Focuses on text representations, not multimodal models.
A Responsibility-Centered Coordination Mechanism for Decentralized Multi-Agent Systems Proposes a decentralized coordination mechanism using commitment tracking to preserve clear accountability across multi-agent delegation chains. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Presents a coordination framework, not a general evaluation benchmark.
System1-Arena: Which Decisions Can Be Left to System 1? Introduces a benchmark evaluating single-forward-pass discrete decision models versus generative reasoning models across diverse closed-loop tasks. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Does not focus on multimodal or vision-language modeling.
System1-Arena: Which Decisions Can Be Left to System 1? Introduces a benchmark evaluating single-forward-pass discrete decision models versus generative reasoning models across diverse closed-loop tasks. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant SWE-bench is merely a subcomponent of a broad decision benchmark.
System1-Arena: Which Decisions Can Be Left to System 1? Introduces a benchmark evaluating single-forward-pass discrete decision models versus generative reasoning models across diverse closed-loop tasks. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Evaluates decision accuracy and latency, not safety guardrails.
System1-Arena: Which Decisions Can Be Left to System 1? Introduces a benchmark evaluating single-forward-pass discrete decision models versus generative reasoning models across diverse closed-loop tasks. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Does not evaluate retrieval-augmented generation architectures.
What do AI chatbots do after users decline therapeutic redirection for suicide thoughts? Audits how persistent conversational agents remain with therapeutic crisis referrals after users repeatedly refuse help for suicidal thoughts. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Studies dialogue response patterns, not autonomous agentic tool use.
Fractional Power-of-Two Quantization for Efficient and Effective Multiplier-Free LLM Inference Proposes a fractional power-of-two quantization grid to enable efficient multiplier-free integer inference for large language models. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on hardware-level quantization, not agent planning.
Fractional Power-of-Two Quantization for Efficient and Effective Multiplier-Free LLM Inference Proposes a fractional power-of-two quantization grid to enable efficient multiplier-free integer inference for large language models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Evaluates existing benchmarks rather than proposing evaluation methodology.
Fractional Power-of-Two Quantization for Efficient and Effective Multiplier-Free LLM Inference Proposes a fractional power-of-two quantization grid to enable efficient multiplier-free integer inference for large language models. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on inference hardware efficiency, not safety risks.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Examines representation interpretability, not model compression or efficiency.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Investigates internal representations of sparse autoencoders, not agents.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Does not evaluate code generation or software engineering.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Does not address retrieval or external knowledge augmentation.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Does not study multimodal models or vision-language tasks.
Behavioral Compositionality of Sparse Autoencoder Representations Introduces a metric to measure how non-linearly features learned by sparse autoencoders interact during downstream model behaviors. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Studies interpretability metrics rather than LLM benchmarking standards.
TRACER: A Multimodal Harness for Patient Care in Hospitals Introduces a multimodal harness and benchmark pairing continuous video with clinical records to assist patient monitoring in hospitals. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:17	Not relevant Applies to hospital patient monitoring, not manufacturing defect detection.
TRACER: A Multimodal Harness for Patient Care in Hospitals Introduces a multimodal harness and benchmark pairing continuous video with clinical records to assist patient monitoring in hospitals. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Evaluates its specific system rather than proposing a general LLM benchmark.
TRACER: A Multimodal Harness for Patient Care in Hospitals Introduces a multimodal harness and benchmark pairing continuous video with clinical records to assist patient monitoring in hospitals. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on monitoring capabilities rather than safety guardrails or alignment.
TRACER: Trajectory-Aware Tool Retrieval for LLM Agents Proposes a trajectory-aware retrieval method that indexes tool sequences in a trie to cut token costs and improve selection accuracy. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Does not focus on code generation or software engineering.
TRACER: Trajectory-Aware Tool Retrieval for LLM Agents Proposes a trajectory-aware retrieval method that indexes tool sequences in a trie to cut token costs and improve selection accuracy. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Presents a tool selection method, not an evaluation benchmark.
TRACER: Trajectory-Aware Tool Retrieval for LLM Agents Proposes a trajectory-aware retrieval method that indexes tool sequences in a trie to cut token costs and improve selection accuracy. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Retrieves tool definitions, not text passages or external documents.
TRACER: Trajectory-Aware Tool Retrieval for LLM Agents Proposes a trajectory-aware retrieval method that indexes tool sequences in a trie to cut token costs and improve selection accuracy. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Does not address safety, guardrails, or alignment risks.
TRACER: Trajectory-Aware Tool Retrieval for LLM Agents Proposes a trajectory-aware retrieval method that indexes tool sequences in a trie to cut token costs and improve selection accuracy. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Operates on text-based tool trajectories, not multimodal inputs.
Auditing Escalation Reliability in Healthcare LLM Agents: A Perturbation-Controlled Benchmark of Workflow-State Failures Introduces an episode-level healthcare benchmark to audit how reliably language model agents escalate actions when facing workflow disruptions. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Does not explore model compression, quantization, or efficient serving.
Auditing Escalation Reliability in Healthcare LLM Agents: A Perturbation-Controlled Benchmark of Workflow-State Failures Introduces an episode-level healthcare benchmark to audit how reliably language model agents escalate actions when facing workflow disruptions. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Focuses on clinical workflows, not software engineering or code.
Auditing Escalation Reliability in Healthcare LLM Agents: A Perturbation-Controlled Benchmark of Workflow-State Failures Introduces an episode-level healthcare benchmark to audit how reliably language model agents escalate actions when facing workflow disruptions. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Evaluates text-based LLM agents, not multimodal models.
Auditing Escalation Reliability in Healthcare LLM Agents: A Perturbation-Controlled Benchmark of Workflow-State Failures Introduces an episode-level healthcare benchmark to audit how reliably language model agents escalate actions when facing workflow disruptions. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Does not study external retrieval or document augmentation.
Diviner: Agentic RCA with Human-Governed Adaptation for Large-Scale Infrastructure Presents an agentic system with human-governed adaptation to automate root-cause analysis across complex, large-scale infrastructure environments. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Focuses on infrastructure incident RCA agents, not benchmark methodology.
Mallea: Building, Steering, and Memorizing Worlds with Coding Agent The paper presents an agent system that generates, edits, and renders interactive virtual worlds from text using executable code rules. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Presents a world simulation system, not benchmark methodology.
Mallea: Building, Steering, and Memorizing Worlds with Coding Agent The paper presents an agent system that generates, edits, and renders interactive virtual worlds from text using executable code rules. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Focuses on code-driven world generation, not multimodal VLM architecture.
Mallea: Building, Steering, and Memorizing Worlds with Coding Agent The paper presents an agent system that generates, edits, and renders interactive virtual worlds from text using executable code rules. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on world generation via code, not safety.
Mallea: Building, Steering, and Memorizing Worlds with Coding Agent The paper presents an agent system that generates, edits, and renders interactive virtual worlds from text using executable code rules. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:17	Not relevant Targets world generation simulation, not industrial defect detection.
The Density of Thought: Predicting and Steering Reasoning in Large Language Models The paper explores whether measuring information density across reasoning steps can predict and guide answer correctness in large language models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Studies chain-of-thought density, not benchmark evaluation design.
The Density of Thought: Predicting and Steering Reasoning in Large Language Models The paper explores whether measuring information density across reasoning steps can predict and guide answer correctness in large language models. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Analyzes reasoning chain information theory, not deployment safety.
The Density of Thought: Predicting and Steering Reasoning in Large Language Models The paper explores whether measuring information density across reasoning steps can predict and guide answer correctness in large language models. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Analyzes internal reasoning density of LLMs, not autonomous agents.
Bias Control and Injection for Multi-Agent Aggregation The paper proposes a test-time aggregation method using bias control and injection to resolve disagreements among reasoning language model agents. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Focuses on collective agent decision aggregation, not model distillation.
Bias Control and Injection for Multi-Agent Aggregation The paper proposes a test-time aggregation method using bias control and injection to resolve disagreements among reasoning language model agents. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes an aggregation technique, not benchmark evaluation methodology.
Bias Control and Injection for Multi-Agent Aggregation The paper proposes a test-time aggregation method using bias control and injection to resolve disagreements among reasoning language model agents. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Applies to text-based reasoning, not multimodal vision models.
Bias Control and Injection for Multi-Agent Aggregation The paper proposes a test-time aggregation method using bias control and injection to resolve disagreements among reasoning language model agents. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Focuses on multi-agent consensus, not retrieval-augmented generation.
Bias Control and Injection for Multi-Agent Aggregation The paper proposes a test-time aggregation method using bias control and injection to resolve disagreements among reasoning language model agents. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Evaluates mathematical reasoning, not software code generation.

1,681–1,740 of 13,754← Previous Next →