{"arxiv_id":"2604.01754","incoming":{"papers":13,"sentences":17,"groups":[{"kind":"data_used","label":"Uses its data or values","papers":3,"sentences":3,"quotes":[{"arxiv_id":"2608.09629","title":"Rethinking Self-Evolving Agents: Do We Still Need Prescribed Optimization Pipelines?","url":"/paper/2608.09629","section":"A frontier optimizer can improve skills without prescribed pipelines","sentence":"The comparison spans the official SearchQA and OfficeQA evaluation splits, SpreadsheetBench, and LiveMathematicianBench (abbreviated LiveMath) [ref] , with Qwen3.5-4B and GPT-5.5 as target models [ref] .","intent_p":0.81,"contests_p":null},{"arxiv_id":"2607.28048","title":"SKILL-KD: Contrastive Skill Distillation for LLM Agents","url":"/paper/2607.28048","section":"Experiments","sentence":"The main experiment reports each benchmark's native hard-success or exact-match score on the test partitions across SearchQA [ref] , SpreadsheetBench [ref] , DocVQA [ref] , LiveMath [ref] , and ALFWorld [ref] .","intent_p":0.89,"contests_p":null},{"arxiv_id":"2608.27454","title":"WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution","url":"/paper/2608.27454","section":"Experimental Setup","sentence":"Datasets We evaluate across five benchmarks spanning diverse domains: mathematical reasoning (LiveMathematicianBench (LiveMath) [ref] ), web search (SealQA [ref] ), spreadsheet manipulation (SpreadsheetBench (SpreadSheet) [ref] ), long-context document question answering OfficeQA [ref] ), and interactive embodied tasks (ALFWorld [ref] ).","intent_p":0.81,"contests_p":null}]}],"unlabelled_sentences":14},"outgoing":{"papers":7,"sentences":7,"groups":[{"kind":"supports_this_paper","label":"Cites as support","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2411.04872","title":"FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI","url":"/paper/2411.04872","section":"Introduction","sentence":"Studies suggest that many problems have leaked into pre-training corpora, allowing models to memorize solution templates rather than derive answers from first principles [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2511.02864","title":"Mathematical exploration and discovery at scale","url":"/paper/2511.02864","section":"Endnote and Discussion","sentence":"This is precisely the kind of ability mathematicians hope AI systems may leverage on when facilitating research, as illustrated by the work of [ref] using AlphaEvolve for mathematical exploration and discovery.","intent_p":0.97,"contests_p":null}]}],"unlabelled_sentences":5},"claims":[],"thresholds":{"intent_p":0.8,"supports_p":0.95,"contests_p":0.7},"method":"Citing sentences come from each paper's LaTeX source; every \\cite marks the sentence and the works it cites. Pith's classifier typed each sentence. A label is shown at intent probability 0.8 or more (0.95 for support), a contest at 0.7 or more. Measured accuracy against two blind model families is at https://pith.science/accuracy. The quoted sentences are the citing authors' own.","held_up":null,"_links":{"html":"https://pith.science/paper/2604.01754"}}