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
Cheap, Consistent, Calibrated, but Biased? Typed Decision Models as Judges The paper evaluates typed decision models against standard language model judges to measure how emotional tone and biases affect their evaluation accuracy. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not examine agents, planning, or tool use.
Cheap, Consistent, Calibrated, but Biased? Typed Decision Models as Judges The paper evaluates typed decision models against standard language model judges to measure how emotional tone and biases affect their evaluation accuracy. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses entirely on text-based decision models.
Cheap, Consistent, Calibrated, but Biased? Typed Decision Models as Judges The paper evaluates typed decision models against standard language model judges to measure how emotional tone and biases affect their evaluation accuracy. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Studies decision model bias, not model compression or efficiency techniques.
Cheap, Consistent, Calibrated, but Biased? Typed Decision Models as Judges The paper evaluates typed decision models against standard language model judges to measure how emotional tone and biases affect their evaluation accuracy. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not cover code models or software engineering.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not focused on code generation or software engineering.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Focuses on inference hardware optimization, not safety or guardrails.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Addresses language model inference, not multimodal architectures.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Does not introduce or study benchmarks for LLM evaluation.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not address retrieval-augmented generation architectures.
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware The paper presents SparPR, a runtime key-value cache recovery mechanism that reduces memory bottlenecks during model inference on consumer hardware. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not study multi-step agents, planning, or tool use.
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness Evaluates AI text detectors under attacks applied to both human and machine text, revealing high false-positive rates on edited human writing. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not cover language model agents or planning.
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness Evaluates AI text detectors under attacks applied to both human and machine text, revealing high false-positive rates on edited human writing. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not involve retrieval-augmented generation systems.
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness Evaluates AI text detectors under attacks applied to both human and machine text, revealing high false-positive rates on edited human writing. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not cover model compression or inference acceleration.
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness Evaluates AI text detectors under attacks applied to both human and machine text, revealing high false-positive rates on edited human writing. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Covers text detection, not multimodal models.
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness Evaluates AI text detectors under attacks applied to both human and machine text, revealing high false-positive rates on edited human writing. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to code generation or software engineering.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Addresses diffusion language models, not multimodal data.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not focus on coding or software engineering.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not use external retrieval or external knowledge bases.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on hallucination mitigation rather than model compression.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Focuses on mitigation methods rather than evaluation benchmarks.
TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models Proposes dynamic activation steering using preceding denoising predictions to reduce hallucinations in diffusion language models. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not involve multi-step agent actions or planning.
SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding Agents Introduces a benchmark evaluating coding agent misbehaviors across code generation, file management, and system environment integrity. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Does not cover vision, video, or multimodal models.
SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding Agents Introduces a benchmark evaluating coding agent misbehaviors across code generation, file management, and system environment integrity. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model efficiency, distillation, or serving.
SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding Agents Introduces a benchmark evaluating coding agent misbehaviors across code generation, file management, and system environment integrity. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not address retrieval-augmented generation architectures.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Does not propose evaluation benchmarks or benchmark analysis.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on model compression, quantization, or serving.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not involve retrieval or document grounding.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Addresses translation quality and expressiveness, not safety or guardrails.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to programming or software engineering.
PACT: A Pragmatic Prompting Framework for Expressive and Voice-Preserving Speech-to-Speech Translation Presents a pragmatic prompting framework to preserve speaker voice, emotion, and tone during speech-to-speech translation. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not cover planning, tool use, or agent workflows.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Does not focus on multimodal models or cross-modal tasks.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not focus on planning, tools, or agents.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not specific to coding or software engineering tasks.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Addresses evaluation redundancy rather than safety or reliability risks.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on benchmark dataset compression, not model compression or serving.
Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models Surveys methods for compressing large language model benchmarks to reduce evaluation costs while maintaining accurate score estimates. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not address retrieval-augmented generation.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses entirely on clinical text without multimodal inputs.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Addresses medical information extraction rather than deployment safety risks.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Introduces a medical domain dataset rather than general LLM evaluation.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not address code models or software engineering.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not focus on autonomous multi-step agents or tools.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not involve retrieval-augmented generation techniques.
PAINT: A Framework for Multilingual Pain Representation from Medical Licensing Exams to Real-World Clinical Texts Presents a multilingual framework and dataset spanning four languages to evaluate clinical pain representations across exams and clinical notes. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not address model compression or inference efficiency.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not cover model compression or inference efficiency.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not involve retrieval-augmented generation systems.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Does not study autonomous agents or tool use.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Focuses on preference evaluation benchmarks rather than deployment safety mechanisms.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to software engineering or code generation.
LegalRewardBench: Benchmarking Reward Models for Legal Preference Evaluation Introduces a human-verified benchmark to evaluate how accurately reward models judge language model outputs on legal tasks. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Focuses exclusively on text evaluation in the legal domain.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	Retrieval-augmented generation (RAG) via arxiv · 15 Oct, 06:17	Not relevant Does not propose novel retrieval-augmented generation methods.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	AI safety and reliability in deployment via arxiv · 15 Oct, 06:17	Not relevant Addresses content presentation optimization, not safety or guardrails.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Does not investigate model efficiency, compression, or quantization.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Uses LLM judges as a component, but does not benchmark LLMs.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Not related to code generation or software engineering.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	AI agents via arxiv · 15 Oct, 06:17	Not relevant Focuses on presentation optimization rather than autonomous agent systems.
Prompt or Policy? Optimizing Physician Slate Presentations with LLM User Simulators Develops a simulation framework using persona-guided models to optimize comparative doctor summary presentations for medical patients. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Deals strictly with text-based summaries and profiles.
Jev-Guard: Fast Routing and Auditing Agent Actions over Editable Policy Trees Introduces a low-latency guardrail model that routes incoming agent actions through editable policy trees for safety auditing. Preprint	Multimodal models via arxiv · 15 Oct, 06:17	Not relevant Does not involve multimodal models or data.
Jev-Guard: Fast Routing and Auditing Agent Actions over Editable Policy Trees Introduces a low-latency guardrail model that routes incoming agent actions through editable policy trees for safety auditing. Preprint	LLMs for code via arxiv · 15 Oct, 06:17	Not relevant Does not evaluate code generation or software engineering assistants.
Jev-Guard: Fast Routing and Auditing Agent Actions over Editable Policy Trees Introduces a low-latency guardrail model that routes incoming agent actions through editable policy trees for safety auditing. Preprint	Small and efficient models via arxiv · 15 Oct, 06:17	Not relevant Focuses on latency for guardrails, not model compression or efficiency methods.
Jev-Guard: Fast Routing and Auditing Agent Actions over Editable Policy Trees Introduces a low-latency guardrail model that routes incoming agent actions through editable policy trees for safety auditing. Preprint	LLM evaluation and benchmarks via arxiv · 15 Oct, 06:17	Not relevant Focuses on deploying guardrail routing rather than benchmark design.

11,461–11,520 of 17,313← Previous Next →