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
Multi-Stream Residual Computation in Diffusion Language Models Equips diffusion language models with multiple residual streams and adaptive routing to separate visible context from masked token generation. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Does not propose evaluation frameworks or benchmark analyses.
Prompt-Injection Probes Learn the Conversation, Not Just the Injection Shows linear prompt-injection monitors overfit to conversation features and reduces their false alarms by adding matched harmless negative examples. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, or efficient serving.
Prompt-Injection Probes Learn the Conversation, Not Just the Injection Shows linear prompt-injection monitors overfit to conversation features and reduces their false alarms by adding matched harmless negative examples. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not focused on software engineering or coding assistants.
Prompt-Injection Probes Learn the Conversation, Not Just the Injection Shows linear prompt-injection monitors overfit to conversation features and reduces their false alarms by adding matched harmless negative examples. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses on text-based prompt injection, not multimodal models.
Prompt-Injection Probes Learn the Conversation, Not Just the Injection Shows linear prompt-injection monitors overfit to conversation features and reduces their false alarms by adding matched harmless negative examples. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Focuses on injection detection rather than retrieval-augmented generation systems.
Prompt-Injection Probes Learn the Conversation, Not Just the Injection Shows linear prompt-injection monitors overfit to conversation features and reduces their false alarms by adding matched harmless negative examples. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Focuses on detection defense mechanisms rather than benchmark development.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Does not address AI safety, jailbreaks, or guardrails.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Focuses on statistical psycholinguistic estimators rather than LLM benchmarking.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, quantization, or efficient serving.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not address software engineering or code generation.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Studies computational psycholinguistics and reading times, not LLM agents.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Deals with text unigram surprisal, not multimodal models.
Revisiting the Estimation of Unigram Surprisal in Computational Psycholinguistics Proposes estimating both contextual and unigram surprisal from the same language model to reduce estimation mismatch in reading time studies. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not address retrieval-augmented generation.
Receiver Decides Agent-to-Agent Negotiations Studies agent-to-agent negotiations over money, finding that the receiving agent's characteristics primarily dictate the final transfer amounts. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not investigate model compression, distillation algorithms, or quantization.
Receiver Decides Agent-to-Agent Negotiations Studies agent-to-agent negotiations over money, finding that the receiving agent's characteristics primarily dictate the final transfer amounts. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not involve retrieval-augmented generation.
Receiver Decides Agent-to-Agent Negotiations Studies agent-to-agent negotiations over money, finding that the receiving agent's characteristics primarily dictate the final transfer amounts. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses on text-based agent negotiations, not multimodal models.
Receiver Decides Agent-to-Agent Negotiations Studies agent-to-agent negotiations over money, finding that the receiving agent's characteristics primarily dictate the final transfer amounts. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not focused on software engineering or code generation.
Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments Presents a randomized experiment framework to test and curate self-written agent playbooks, preventing task failure caused by unverified lessons. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not study model compression, quantization, or efficient serving.
Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments Presents a randomized experiment framework to test and curate self-written agent playbooks, preventing task failure caused by unverified lessons. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Focuses on persistent textual memory and playbooks for agents, not RAG.
Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments Presents a randomized experiment framework to test and curate self-written agent playbooks, preventing task failure caused by unverified lessons. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not specifically focused on code generation or software repair.
Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments Presents a randomized experiment framework to test and curate self-written agent playbooks, preventing task failure caused by unverified lessons. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Presents an agent memory curation method rather than an evaluation benchmark.
Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments Presents a randomized experiment framework to test and curate self-written agent playbooks, preventing task failure caused by unverified lessons. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses on agent textual memory, not multimodal systems.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not address retrieval-augmented generation systems.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Focuses on cognitive linguistics and speech perception, not deployment safety.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to code generation or software engineering.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Studies psycholinguistic surprisal from speech, not autonomous agents.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not cover model compression, distillation, or efficient inference.
Estimating Surprisal from the Speech Signal Estimates surprisal directly from continuous speech signals using a speech language model and importance sampling over abstract linguistic units. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Presents an estimator for psycholinguistic surprisal, not LLM benchmark evaluation.
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit Identifies passage copying as a failure mode in long-form retrieval systems and introduces sentence-level credit assignment to encourage genuine grounding. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not address code generation or software engineering.
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit Identifies passage copying as a failure mode in long-form retrieval systems and introduces sentence-level credit assignment to encourage genuine grounding. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression or efficient serving techniques.
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit Identifies passage copying as a failure mode in long-form retrieval systems and introduces sentence-level credit assignment to encourage genuine grounding. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Focuses on long-form RAG training objectives rather than autonomous agents.
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit Identifies passage copying as a failure mode in long-form retrieval systems and introduces sentence-level credit assignment to encourage genuine grounding. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Proposes a training method for RAG, not a general benchmark.
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit Identifies passage copying as a failure mode in long-form retrieval systems and introduces sentence-level credit assignment to encourage genuine grounding. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Deals with textual long-form RAG, not multimodal models.
Teacher-Guided Tool Learning for Small Language Models via Bidirectional KL Constraints Proposes a teacher-guided reinforcement learning method using bidirectional constraints to help small language models learn effective tool use. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Focuses on tool-use learning rather than retrieval-augmented generation.
Teacher-Guided Tool Learning for Small Language Models via Bidirectional KL Constraints Proposes a teacher-guided reinforcement learning method using bidirectional constraints to help small language models learn effective tool use. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Evaluates tool use on API benchmarks, not software engineering or code generation.
Teacher-Guided Tool Learning for Small Language Models via Bidirectional KL Constraints Proposes a teacher-guided reinforcement learning method using bidirectional constraints to help small language models learn effective tool use. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Does not involve multimodal or vision-language models.
Teacher-Guided Tool Learning for Small Language Models via Bidirectional KL Constraints Proposes a teacher-guided reinforcement learning method using bidirectional constraints to help small language models learn effective tool use. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Introduces a training algorithm, not an evaluation benchmark or judge framework.
Teacher-Guided Tool Learning for Small Language Models via Bidirectional KL Constraints Proposes a teacher-guided reinforcement learning method using bidirectional constraints to help small language models learn effective tool use. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Focuses on tool capability acquisition rather than AI safety or guardrails.
How Safe Is ChatGPT Health Advice? Evidence from Real User Conversations in Four Global South Countries Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not investigate retrieval-augmented generation architectures.
How Safe Is ChatGPT Health Advice? Evidence from Real User Conversations in Four Global South Countries Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Examines text health conversations, not multimodal models.
How Safe Is ChatGPT Health Advice? Evidence from Real User Conversations in Four Global South Countries Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to software engineering or code generation.
How Safe Is ChatGPT Health Advice? Evidence from Real User Conversations in Four Global South Countries Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Studies conversational health advice, not autonomous multi-step agents.
How Safe Is ChatGPT Health Advice? Evidence from Real User Conversations in Four Global South Countries Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression, pruning, or efficient inference.
The Last Mile of Coding Agents for Long-Document Tasks: An Empirical Study on Clinical Trial Protocol Digitization Introduces a benchmark evaluating how reliably software coding agents extract and digitize complex unstructured clinical trial protocol documents. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Evaluates coding agents on documents, not external RAG architectures.
The Last Mile of Coding Agents for Long-Document Tasks: An Empirical Study on Clinical Trial Protocol Digitization Introduces a benchmark evaluating how reliably software coding agents extract and digitize complex unstructured clinical trial protocol documents. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on model compression or efficient serving.
The Last Mile of Coding Agents for Long-Document Tasks: An Empirical Study on Clinical Trial Protocol Digitization Introduces a benchmark evaluating how reliably software coding agents extract and digitize complex unstructured clinical trial protocol documents. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Analyzes agent task performance and cost, not safety, jailbreaks, or guardrails.
The Last Mile of Coding Agents for Long-Document Tasks: An Empirical Study on Clinical Trial Protocol Digitization Introduces a benchmark evaluating how reliably software coding agents extract and digitize complex unstructured clinical trial protocol documents. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses on coding agents processing PDFs into structured data, not multimodal architectures.
Imprinting: Knowledge Distillation Can Shift Student Behavior Demonstrates that knowledge distillation transfers safety dispositions and behavioral tendencies from teacher models to student models on unrelated tasks. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Studies distillation behavior transfer rather than multi-step agent planning.
Imprinting: Knowledge Distillation Can Shift Student Behavior Demonstrates that knowledge distillation transfers safety dispositions and behavioral tendencies from teacher models to student models on unrelated tasks. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Examines text language model distillation, not multimodal models.
Imprinting: Knowledge Distillation Can Shift Student Behavior Demonstrates that knowledge distillation transfers safety dispositions and behavioral tendencies from teacher models to student models on unrelated tasks. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Analyzes distillation side-effects rather than benchmarking or evaluation metrics.
Imprinting: Knowledge Distillation Can Shift Student Behavior Demonstrates that knowledge distillation transfers safety dispositions and behavioral tendencies from teacher models to student models on unrelated tasks. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to code generation or software engineering.
Imprinting: Knowledge Distillation Can Shift Student Behavior Demonstrates that knowledge distillation transfers safety dispositions and behavioral tendencies from teacher models to student models on unrelated tasks. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not cover retrieval-augmented generation.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses on text embedding retrieval across languages, not multimodal data.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Focuses on embedding spaces for dense retrieval rather than agent systems.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Focuses on cross-lingual retrieval accuracy rather than safety or guardrails.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not focus on software engineering or code LLMs.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Proposes a steering method for retrieval, not an evaluation framework.
Retrieval Disagreement as a Signal: Diagnosing and Steering Multilingual Dense Embedders Analyzes retrieval divergence across languages in multilingual dense embedders and introduces disagreement-guided steering to align shared-index search results. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model efficiency, distillation, or quantization.
How Reliable Are System-One Decisions in Medicine? A Comparative Study of Jev, Jev-like Models, and LLMs Evaluates fast, system-one decision models and language models across fifteen medical datasets, testing their accuracy, format robustness, and abstention awareness. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to software development or code generation.
How Reliable Are System-One Decisions in Medicine? A Comparative Study of Jev, Jev-like Models, and LLMs Evaluates fast, system-one decision models and language models across fifteen medical datasets, testing their accuracy, format robustness, and abstention awareness. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Analyzes text-based medical decision tasks, not multimodal models.

8,041–8,100 of 13,754← Previous Next →