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
CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation Proposes a chunk-level credit assignment method for on-policy distillation across language models with different tokenizers. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not relate to retrieval-augmented generation.
CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation Proposes a chunk-level credit assignment method for on-policy distillation across language models with different tokenizers. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Does not study deployment safety, jailbreaks, or guardrails.
CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation Proposes a chunk-level credit assignment method for on-policy distillation across language models with different tokenizers. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not focus on multimodal architectures.
CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation Proposes a chunk-level credit assignment method for on-policy distillation across language models with different tokenizers. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents a distillation method rather than an evaluation benchmark.
CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation Proposes a chunk-level credit assignment method for on-policy distillation across language models with different tokenizers. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Proposes a policy distillation algorithm, not autonomous agent planning.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Does not investigate safety risks, prompt injections, or red-teaming.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on retrieval mechanisms rather than multi-step agent systems.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents a retrieval method rather than a novel LLM evaluation benchmark.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not involve multimodal models or vision-language tasks.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Focuses on retriever ranking performance rather than model compression.
Beyond the Final Embedding: Retrieving Memories from Token-Level Evidence Introduces a retrieval method combining token-level matching and document-level embeddings for memory-augmented language models. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not address code generation or software development.
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study Evaluates whether automated prompt optimizers inadvertently cause regressions in unscored prompt constraints and requirements. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Does not study model compression, distillation, or serving.
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study Evaluates whether automated prompt optimizers inadvertently cause regressions in unscored prompt constraints and requirements. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not focus on multimodal models.
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study Evaluates whether automated prompt optimizers inadvertently cause regressions in unscored prompt constraints and requirements. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not evaluate code-specific models or software engineering tasks.
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study Evaluates whether automated prompt optimizers inadvertently cause regressions in unscored prompt constraints and requirements. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not focus on retrieval-augmented generation.
Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study Evaluates whether automated prompt optimizers inadvertently cause regressions in unscored prompt constraints and requirements. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Analyzes automated prompt optimizers rather than multi-step agent architectures.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not involve code generation or software engineering.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Studies data collection strategies rather than LLM benchmarking or evaluation.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Covers pretraining data collection rather than inference-time retrieval systems.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Does not evaluate model safety, jailbreaks, or deployment guardrails.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on web crawling for pretraining, not autonomous agents.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not examine multimodal architectures or datasets.
Host-Aware Web Crawling for LLM Pretraining Data Studies host-level budget allocation and crawling policies to collect pretraining web data for language models. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Does not focus on model compression, quantization, or efficient inference.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Deals with text tokenization systems rather than multimodal architectures.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Addresses tokenization serving infrastructure, not agent decision-making or workflows.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Does not propose evaluation frameworks or benchmarks.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on serving throughput, not safety or guardrails.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not address code generation or software engineering LLMs.
Taking Tokenization off the Host for Agentic LLM Serving Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Focuses on GPU tokenization rather than RAG or retrieval.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not involve retrieval-augmented generation.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Studies attention mechanisms in text-to-image models, not agent planning.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Does not address deployment safety, guardrails, or jailbreaks.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Focuses on attention dynamics in generative models rather than compression or serving.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not address code generation or software tooling.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Investigates attention mechanisms rather than LLM benchmarking techniques.
Exploring Attention Sinks in Text-to-Image Transformers from Diffusion and Autoregressive Models Analyzes the emergence and behavior of attention sinks in text-to-image diffusion and autoregressive Transformers. Preprint	Computer vision for industrial applications via arxiv · 12 Oct, 16:28	Not relevant Studies foundational attention mechanics in generative models, not industrial vision tasks.
Test-Time Memory Construction Presents a framework that defers memory construction to test time to answer queries across long interaction histories. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Addresses memory retrieval, not deployment safety or adversarial robustness.
Test-Time Memory Construction Presents a framework that defers memory construction to test time to answer queries across long interaction histories. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not involve multimodal representations or models.
Test-Time Memory Construction Presents a framework that defers memory construction to test time to answer queries across long interaction histories. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Focuses on memory management rather than small models, quantization, or pruning.
Test-Time Memory Construction Presents a framework that defers memory construction to test time to answer queries across long interaction histories. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Does not focus on code synthesis or software engineering.
Test-Time Memory Construction Presents a framework that defers memory construction to test time to answer queries across long interaction histories. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Presents a memory retrieval architecture rather than LLM benchmarking methods.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Does not study deployment safety, jailbreaks, or alignment.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Evaluates MoE routing across generic categories, not code-specific LLM systems.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Proposes an architectural routing bias rather than evaluation methodologies.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Addresses token routing in MoE models generally, not multimodal architectures.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Addresses routing in Mixture-of-Experts architectures, not agent systems.
Overcoming Data Imbalances with Expert Selection Biases Proposes category-conditioned expert routing biases to mitigate data imbalance in mixture-of-experts language models. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not involve retrieval-augmented generation.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Does not focus on multimodal inputs or models.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Focuses on evaluation rubrics for models rather than autonomous agents.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Does not investigate model compression, distillation, or efficient inference.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Focuses on evaluation rubric structures rather than safety risks or guardrails.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Not focused on code generation or software engineering tools.
From Structure to Validity: Mechanism-Guided Generation and Refinement of LLM Rubrics Analyzes rubric construction mechanisms and introduces a tree-search method to refine rubrics for evaluating language model outputs. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not cover retrieval-augmented generation systems.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	AI agents via arxiv · 12 Oct, 16:28	Not relevant Pertains to prompt optimization for classification, not multi-step agent planning.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	Small and efficient models via arxiv · 12 Oct, 16:28	Not relevant Focuses on prompt refinement rather than model compression or efficient serving.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	Retrieval-augmented generation (RAG) via arxiv · 12 Oct, 16:28	Not relevant Does not involve external document retrieval or grounding.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	AI safety and reliability in deployment via arxiv · 12 Oct, 16:28	Not relevant Targets stance detection classification, not safety guardrails or red-teaming.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	Multimodal models via arxiv · 12 Oct, 16:28	Not relevant Deals with text stance detection without multimodal data.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	LLMs for code via arxiv · 12 Oct, 16:28	Not relevant Not related to code generation or software engineering.
MoPrO-SD: Modular Prompt Optimization for Stance Detection Presents a modular prompt optimization framework and benchmark targeting stance detection in text. Preprint	LLM evaluation and benchmarks via arxiv · 12 Oct, 16:28	Not relevant Proposes a prompt optimization method rather than general benchmark evaluation frameworks.

12,601–12,660 of 13,754← Previous Next →