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
Following Delegation Traces to Study Human Oversight in Interaction with Multimodal AI Investigates how user trust influences delegation and verification choices during human interaction with multimodal video question-answering systems. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Human-subject interaction study rather than an LLM evaluation benchmark.
Following Delegation Traces to Study Human Oversight in Interaction with Multimodal AI Investigates how user trust influences delegation and verification choices during human interaction with multimodal video question-answering systems. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:17	Not relevant Human-AI interaction study, not industrial defect detection.
Why Shallow Neural Networks Learn Sinusoidal Components First: An Orthogonal Polynomial Dynamics Framework Presents a theoretical orthogonal polynomial framework explaining why shallow neural networks learn low-frequency sinusoidal components before high frequencies. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Pure theoretical learning dynamics, not deployment safety or reliability.
Why Shallow Neural Networks Learn Sinusoidal Components First: An Orthogonal Polynomial Dynamics Framework Presents a theoretical orthogonal polynomial framework explaining why shallow neural networks learn low-frequency sinusoidal components before high frequencies. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Theoretical neural network spectral bias study, not about AI agents.
Why Shallow Neural Networks Learn Sinusoidal Components First: An Orthogonal Polynomial Dynamics Framework Presents a theoretical orthogonal polynomial framework explaining why shallow neural networks learn low-frequency sinusoidal components before high frequencies. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Theoretical analysis of shallow networks, not model compression or efficiency.
Why Shallow Neural Networks Learn Sinusoidal Components First: An Orthogonal Polynomial Dynamics Framework Presents a theoretical orthogonal polynomial framework explaining why shallow neural networks learn low-frequency sinusoidal components before high frequencies. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Focuses on mathematical theory of gradient descent, not LLM benchmarks.
Time to Fake It: Continual Machine Learning Intensifies Unfairness in Deepfake Detection Analyzes how continual learning methods affect demographic fairness over time across eleven deepfake generation techniques. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on deepfake detection fairness, not language model agents or planning.
Time to Fake It: Continual Machine Learning Intensifies Unfairness in Deepfake Detection Analyzes how continual learning methods affect demographic fairness over time across eleven deepfake generation techniques. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Studies deepfake classifier fairness over time rather than LLM benchmarking or evaluation.
Time to Fake It: Continual Machine Learning Intensifies Unfairness in Deepfake Detection Analyzes how continual learning methods affect demographic fairness over time across eleven deepfake generation techniques. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Does not study model compression, distillation, quantization, or efficient model serving.
OntoAct: Hypothesis-Based Neuro-Symbolic Agents through Ontological Abstraction The paper pairs language models with symbolic verifiers to test and update reusable hypotheses for environment diagnostics and repair. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Develops a neuro-symbolic agent framework rather than safety alignment, guardrails, or red-teaming.
OntoAct: Hypothesis-Based Neuro-Symbolic Agents through Ontological Abstraction The paper pairs language models with symbolic verifiers to test and update reusable hypotheses for environment diagnostics and repair. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Focuses on neuro-symbolic agent architecture rather than model compression or inference efficiency.
GraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal Cancers The paper aggregates irregular electronic health records into temporal graphs to predict risks of gastrointestinal cancers. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Does not address model compression, quantization, distillation, or efficient model serving.
GraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal Cancers The paper aggregates irregular electronic health records into temporal graphs to predict risks of gastrointestinal cancers. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on clinical temporal graph learning, not LLM agents or multi-step tool use.
GraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal Cancers The paper aggregates irregular electronic health records into temporal graphs to predict risks of gastrointestinal cancers. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Applies GNNs to healthcare records rather than AI deployment safety, jailbreaks, or guardrails.
GraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal Cancers The paper aggregates irregular electronic health records into temporal graphs to predict risks of gastrointestinal cancers. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes a clinical risk prediction model rather than LLM benchmarking or evaluation.
CARA: Causally Guided Retrieval Agent for Question Answering The paper guides knowledge-graph retrieval agents by evaluating the causal impact of evidence on answering questions rather than text similarity. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes a retrieval system rather than general LLM evaluation frameworks or benchmarks.
CARA: Causally Guided Retrieval Agent for Question Answering The paper guides knowledge-graph retrieval agents by evaluating the causal impact of evidence on answering questions rather than text similarity. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on retrieval precision and question answering rather than deployment safety or guardrails.
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe The paper outlines a long-context midtraining stage using software engineering data to prepare base language models for agentic tasks. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Presents a large-scale training recipe rather than model compression, distillation, or quantization.
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe The paper outlines a long-context midtraining stage using software engineering data to prepare base language models for agentic tasks. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Addresses text-based code and agent traces, not multimodal vision or audio processing.
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe The paper outlines a long-context midtraining stage using software engineering data to prepare base language models for agentic tasks. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Focuses on midtraining models on software data, not retrieval-augmented generation architectures.
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe The paper outlines a long-context midtraining stage using software engineering data to prepare base language models for agentic tasks. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Evaluates on existing SWE benchmarks rather than introducing new evaluation methods or benchmarks.
Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe The paper outlines a long-context midtraining stage using software engineering data to prepare base language models for agentic tasks. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on agentic code capability midtraining rather than safety alignment or guardrails.
DemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from Demonstration The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Optimizes robotic planning routines rather than neural network compression or efficient model serving.
DemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from Demonstration The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:17	Not relevant Focuses on robotic trajectory generation rather than industrial visual inspection or defect detection.
DemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from Demonstration The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Focuses on robotic demonstration generation rather than AI safety guardrails or LLM reliability.
DemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from Demonstration The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on robotics motion planning rather than language model planning and agents.
DemoStream: Batched Task-and-Motion-Plan Optimization for Automated Learning from Demonstration The paper optimizes batched task and motion planning to generate high-quality autonomous demonstrations for robotic manipulation training. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Addresses robotic motion planning rather than language model benchmarks or evaluation methods.
Selection Is Not Learning: Choosing Model Futures The paper separates task generation from model update inheritance to avoid compounding errors in self-evolving language model loops. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Focuses on self-training algorithms rather than model compression, quantization, or efficiency.
Selection Is Not Learning: Choosing Model Futures The paper separates task generation from model update inheritance to avoid compounding errors in self-evolving language model loops. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Studies general self-training dynamics rather than code generation or software maintenance.
Selection Is Not Learning: Choosing Model Futures The paper separates task generation from model update inheritance to avoid compounding errors in self-evolving language model loops. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Studies language model self-evolution, not multimodal vision-language architectures.
Selection Is Not Learning: Choosing Model Futures The paper separates task generation from model update inheritance to avoid compounding errors in self-evolving language model loops. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Addresses self-evolving model training loops rather than retrieval-augmented generation systems.
Selection Is Not Learning: Choosing Model Futures The paper separates task generation from model update inheritance to avoid compounding errors in self-evolving language model loops. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on model weight updating loops, not multi-step tool-using agents.
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts The paper tests how language models weigh price momentum, commentary tone, and earnings surprises when forecasting financial outcomes. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Studies prompt reliance across model sizes, not model compression or efficient inference methods.
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts The paper tests how language models weigh price momentum, commentary tone, and earnings surprises when forecasting financial outcomes. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Analyzes financial prompt reliance, not tool use or autonomous multi-step agents.
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts The paper tests how language models weigh price momentum, commentary tone, and earnings surprises when forecasting financial outcomes. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Focuses on prompt bias in financial forecasting rather than retrieval-augmented generation architectures.
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts The paper tests how language models weigh price momentum, commentary tone, and earnings surprises when forecasting financial outcomes. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Examines text-based financial prompts, not multimodal image, video, or audio tasks.
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts The paper tests how language models weigh price momentum, commentary tone, and earnings surprises when forecasting financial outcomes. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Investigates financial numerical forecasting rather than code generation or software engineering.
Decoding Naturalistic Speech from fMRI with Temporal Modeling and Multimodal Alignment The paper aligns functional brain scans with multimodal representations to decode and retrieve perceived natural speech segments. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on brain fMRI decoding, not LLM agent systems.
Decoding Naturalistic Speech from fMRI with Temporal Modeling and Multimodal Alignment The paper aligns functional brain scans with multimodal representations to decode and retrieve perceived natural speech segments. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Applies to neuroscience fMRI decoding rather than LLM benchmarks or evaluation.
Decoding Naturalistic Speech from fMRI with Temporal Modeling and Multimodal Alignment The paper aligns functional brain scans with multimodal representations to decode and retrieve perceived natural speech segments. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Does not address model compression, quantization, distillation, or efficient model serving.
Decoding Naturalistic Speech from fMRI with Temporal Modeling and Multimodal Alignment The paper aligns functional brain scans with multimodal representations to decode and retrieve perceived natural speech segments. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Covers neural speech decoding rather than AI deployment safety, guardrails, or hallucination.
Progress Reward Model for Reinforcement Learning via Large Language Models The paper uses language model subtask decomposition and dense progress tracking to enhance reinforcement learning on long-horizon tasks. Peer-reviewed	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Addresses reinforcement learning reward shaping rather than AI safety or reliability safeguards.
Progress Reward Model for Reinforcement Learning via Large Language Models The paper uses language model subtask decomposition and dense progress tracking to enhance reinforcement learning on long-horizon tasks. Peer-reviewed	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Focuses on reward shaping for reinforcement learning, not LLM benchmarking or evaluation.
FAOPs: Normal Forms and an Algebra for AI Agent Workflows The paper provides a formal algebraic framework and normal forms to structure and audit enterprise AI agent workflows. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes workflow normalization theory rather than benchmarking or evaluating LLM capabilities.
FAOPs: Normal Forms and an Algebra for AI Agent Workflows The paper provides a formal algebraic framework and normal forms to structure and audit enterprise AI agent workflows. Preprint	LLMs for code via arxiv · 17 Oct, 06:17	Not relevant Focuses on agent workflow architecture rather than code generation or program repair models.
Prism-1: Position- and Rival-Independent Scoring for System One Models The paper evaluates single-pass decision models against rational choice axioms and identifies ordering and rival-option sensitivity failures. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Analyzes single-step choice scoring rather than multi-step tool-using agent workflows.
When Reasoning Is Not Enough: Proactive Inquiry for Cross-Modal Safety in Multimodal Large Language Models The paper enables multimodal models to proactively ask clarifying questions when ambiguous inputs carry cross-modal safety risks. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes a safety inquiry framework rather than new general LLM evaluation methodologies.
When Reasoning Is Not Enough: Proactive Inquiry for Cross-Modal Safety in Multimodal Large Language Models The paper enables multimodal models to proactively ask clarifying questions when ambiguous inputs carry cross-modal safety risks. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on multimodal safety refusal, not multi-step tool-using agents.
Memory as Weights: Internalizing Long-Term History for Streaming Videos The paper compresses video history into evolving lightweight model weights instead of competing for limited input context length. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Focuses on streaming video memory, not tool use or autonomous agents.
Memory as Weights: Internalizing Long-Term History for Streaming Videos The paper compresses video history into evolving lightweight model weights instead of competing for limited input context length. Preprint	Computer vision for industrial applications via arxiv · 17 Oct, 06:17	Not relevant Addresses streaming video-language models, not industrial defect detection or manufacturing inspection.
Memory as Weights: Internalizing Long-Term History for Streaming Videos The paper compresses video history into evolving lightweight model weights instead of competing for limited input context length. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Proposes a video memory architecture rather than new general LLM evaluation benchmarks.
Memory as Weights: Internalizing Long-Term History for Streaming Videos The paper compresses video history into evolving lightweight model weights instead of competing for limited input context length. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Does not address safety, jailbreaks, robustness, or deployment monitoring.
Learning from Unverified Peers via Shared Residual Distillation The paper develops a self-distillation approach enabling models to learn from their own unverified alternative reasoning paths without ground-truth answers. Preprint	AI safety and reliability in deployment via arxiv · 17 Oct, 06:17	Not relevant Addresses training via self-distillation rather than safety, red-teaming, or guardrails.
Learning from Unverified Peers via Shared Residual Distillation The paper develops a self-distillation approach enabling models to learn from their own unverified alternative reasoning paths without ground-truth answers. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Studies self-distillation training methods, not autonomous tool-use or multi-step agents.
Learning from Unverified Peers via Shared Residual Distillation The paper develops a self-distillation approach enabling models to learn from their own unverified alternative reasoning paths without ground-truth answers. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Introduces a distillation training technique rather than testing or benchmarking methodologies.
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following The paper introduces a commercially permissive multilingual instruction dataset covering five European languages to train open language models. Preprint	LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17	Not relevant Contributes an instruction-tuning dataset rather than proposing new evaluation methodologies or benchmarks.
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following The paper introduces a commercially permissive multilingual instruction dataset covering five European languages to train open language models. Preprint	Multimodal models via arxiv · 17 Oct, 06:17	Not relevant Exclusively focuses on text instruction datasets across European languages, not multimodality.
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following The paper introduces a commercially permissive multilingual instruction dataset covering five European languages to train open language models. Preprint	Small and efficient models via arxiv · 17 Oct, 06:17	Not relevant Focuses on instruction data creation rather than model compression, quantization, or distillation.
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following The paper introduces a commercially permissive multilingual instruction dataset covering five European languages to train open language models. Preprint	Retrieval-augmented generation (RAG) via arxiv · 17 Oct, 06:17	Not relevant Focuses on instruction-tuning datasets, not retrieval-augmented generation or external document grounding.
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following The paper introduces a commercially permissive multilingual instruction dataset covering five European languages to train open language models. Preprint	AI agents via arxiv · 17 Oct, 06:17	Not relevant Presents an instruction dataset, not multi-step tool-using autonomous agent architectures.

2,041–2,100 of 13,754← Previous Next →