202609 hot_news 2026-09-22 hot_news 2026-09-21 hot_news 2026-09-20 hot_news 2026-09-19 hot_news 2026-09-18 hot_news 2026-09-17 hot_news 2026-09-16 hot_news 2026-09-15 hot_news 2026-09-14 hot_news 2026-09-13 hot_news 2026-09-12 hot_news 2026-09-11 hot_news 2026-09-10 hot_news 2026-09-09 hot_news 2026-09-08 hot_news 2026-09-07 hot_news 2026-09-06 hot_news 2026-09-05 hot_news 2026-09-04 hot_news 2026-09-03 hot_news 2026-09-02 hot_news 2026-09-01 202608 hot_news 2026-08-31 hot_news 2026-08-30 hot_n ...

数据来源：ArXiv Domain LLM Domain Papers1. Cost-Effective Automated Judging of Natural-Language Mathematical ProofsAbstract:Grading natural-language mathematical proofs is a recurring cost in evaluating math-reasoning systems, and frontier LLM judges are expensive. We ask whether cheap open-weight models can serve as reliable judges given a candidate proof, a ground-truth proof, and a human-grading rubric. On a 200-instance validation sample of IMO-GradingBench, three cheap judges (GPT-OSS 120B, De ...

数据来源：ArXiv Domain LLM Domain Papers1. Simulator-Grounded Large Language Models for Industrial Causal Reasoning: Tool-Use, Structured Injection, and Plant-Portable Retrieval for Wastewater Treatment Decision SupportAbstract:Wastewater operators need answers grounded in how their plant’s variables interact and how fast effects propagate, not in generic pretraining text, when asking causal questions such as “why is N2O rising?” or “what happens if I cut aeration by 20%?”. We compare three concret ...

数据来源：ArXiv Domain LLM Domain Papers1. Does a Language Server Save Tokens for Coding Agents? A Measurement Methodology and Preliminary StudyAbstract:Coding agents spend most of their context budget on retrieval. Lexical retrieval (grep) is universal, instant, and zero-setup, but noisy: it cannot tell a definition from a call from a comment. Semantic retrieval via the Language Server Protocol (LSP) is precise and typed, but needs a running, indexed server and pays a per-symbol round-trip. The cl ...

数据来源：ArXiv Domain LLM Domain Papers1. Beyond Raw Transcripts: Structured Persona Extraction for LLM-Based Digital TwinsAbstract:LLM-based “digital twins” aim to simulate how an individual would behavein new environments or respond to novel questions, given some representation of that individual’s prior responses. A common approach constructs this representation from survey transcripts or summaries responses. Prior work shows that compressing long transcripts into shorter LLM-generated summarie ...

数据来源：ArXiv Domain LLM Domain Papers1. Distinguishing Revision and Delayed Elaboration in Incremental Narrative InterpretationAbstract:Both human and AI systems that process narrative or long-form content operate incrementally: input is received over time, and internal representations must be updated accordingly. Incremental interpretation, therefore, depends not only on what is represented but also on how the representational state evolves under new evidence. We distinguish two structurally di ...

数据来源：ArXiv Domain LLM Domain Papers1. Accelerating LLM Inference via Vector Index Based Output EmbeddingsAbstract:Large output embedding matrices create a significant memory bandwidth bottleneck during autoregressive decoding, especially for compact LLMs with large multilingual vocabularies. We reformulate the output projection followed by top-k token selection as a maximum inner product search over token embeddings and replace the dense vocabulary projection with an HNSW-based vector index. T ...

数据来源：ArXiv Domain LLM Domain Papers1. NLP-Driven Knowledge Extraction and Thematic Classification of Translated Ancient Indian Medical TextsAbstract:Ancient Indian medical texts like Sushruta Samhita have extensive information on diseases, treatments, and surgical techniques. Yet, their ancient format and use of intricate vocabulary pose difficulties in accessibility and systematic ordering. The research here utilizes Natural Language Processing (NLP) methods like Named Entity Recognition (NER ...

数据来源：ArXiv Domain LLM Domain Papers1. Behaviorally Grounded User Profiles from the Wild for Personalized Alignment and Multi-Perspective ReasoningAbstract:Persona-driven techniques increasingly adapt large language models (LLMs) to diverse contexts. However, existing methods predominantly rely on rigid, synthetic personas that flatten individual variation, rely on stereotypes, and miss the nuanced signals driving actual human preferences. We introduce profile behavioral grounding, a framework ...

数据来源：ArXiv Domain LLM Domain Papers1. PRO-Step: Step-level Process Reward Optimization for Retrieval-Augmented GenerationAbstract:Retrieval-Augmented Generation enhances Large Language Models by grounding responses in external knowledge, but multi-hop reasoning remains vulnerable to error propagation, where early retrieval failures confound subsequent steps. Standard outcome-based optimization only rewards the final answer, leaving intermediate retrieval and reasoning errors undetected. While ...

数据来源：ArXiv Domain LLM Domain Papers1. Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM AgentsAbstract:A growing body of work improves frozen large language models (LLMs) as agents by evolving their harness: the textual scaffolding around the model, including persona, strategy, format rules, and control heuristics. Existing reflective prompt-evolution methods usually optimize this harness as one flat string. We instead ask where the ...

数据来源：ArXiv Domain LLM Domain Papers1. Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM AgentsAbstract:A growing body of work improves frozen large language models (LLMs) as agents by evolving their harness: the textual scaffolding around the model, including persona, strategy, format rules, and control heuristics. Existing reflective prompt-evolution methods usually optimize this harness as one flat string. We instead ask where the ...

数据来源：ArXiv Domain LLM Domain Papers1. Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM AgentsAbstract:A growing body of work improves frozen large language models (LLMs) as agents by evolving their harness: the textual scaffolding around the model, including persona, strategy, format rules, and control heuristics. Existing reflective prompt-evolution methods usually optimize this harness as one flat string. We instead ask where the ...

数据来源：ArXiv Domain LLM Domain Papers1. How Much Does Corpus Choice Change Dependency-Distance Estimates?Abstract:Dependency-distance estimates derived from a single corpus are routinely treated as properties of a language, yet this assumption has not been tested across independently compiled corpora. We compared mean dependency-distance estimates across 38 same-language treebank pairs from Universal Dependencies v2.18, using concordance correlation, Bland-Altman analysis, and a twelve-specifica ...

数据来源：ArXiv Domain LLM Domain Papers1. How Much Does Corpus Choice Change Dependency-Distance Estimates?Abstract:Dependency-distance estimates derived from a single corpus are routinely treated as properties of a language, yet this assumption has not been tested across independently compiled corpora. We compared mean dependency-distance estimates across 38 same-language treebank pairs from Universal Dependencies v2.18, using concordance correlation, Bland-Altman analysis, and a twelve-specifica ...