{"arxiv_id":"2608.22533","incoming":{"papers":0,"sentences":0,"groups":[],"unlabelled_sentences":0},"outgoing":{"papers":16,"sentences":31,"groups":[{"kind":"method_or_tool","label":"Uses the method or tool of","papers":2,"sentences":2,"quotes":[{"arxiv_id":"2509.17158","title":null,"url":"/paper/2509.17158","section":"Experimental Setup","sentence":"We evaluate on /, the benchmark executed in the Agents Research Environments () runtime [ref] , across five abilities (Execution, Search, Ambiguity, Adaptability, Time), with 40 reference and 40 held-out tasks per ability.","intent_p":0.88,"contests_p":null},{"arxiv_id":"2602.11964","title":null,"url":"/paper/2602.11964","section":"Experimental Setup","sentence":"We evaluate on /, the benchmark executed in the Agents Research Environments () runtime [ref] , across five abilities (Execution, Search, Ambiguity, Adaptability, Time), with 40 reference and 40 held-out tasks per ability.","intent_p":0.88,"contests_p":null}]},{"kind":"comparison","label":"Compares with","papers":3,"sentences":5,"quotes":[{"arxiv_id":"2409.07429","title":"Agent Workflow Memory","url":"/paper/2409.07429","section":"Experimental Setup","sentence":"To compare against published agent-memory baselines, we evaluate offline Agent Workflow Memory (AWM) [ref] , ACE [ref] , and raw trajectory retrieval on GPT-5.5 over Execution, Search, and Ambiguity, testing whether the gains reduce to trajectory replay, induced workflows, or an evolving playbook.","intent_p":0.98,"contests_p":null},{"arxiv_id":"2510.04618","title":"Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models","url":"/paper/2510.04618","section":"Experimental Setup","sentence":"To compare against published agent-memory baselines, we evaluate offline Agent Workflow Memory (AWM) [ref] , ACE [ref] , and raw trajectory retrieval on GPT-5.5 over Execution, Search, and Ambiguity, testing whether the gains reduce to trajectory replay, induced workflows, or an evolving playbook.","intent_p":0.98,"contests_p":null},{"arxiv_id":"2603.23234","title":"MemCollab: Cross-Model Memory Collaboration via Contrastive Trajectory Distillation","url":"/paper/2603.23234","section":"Related Work","sentence":"Most directly related is MemCollab, which likewise contrasts trajectories from different models on the same task [ref] —but toward transferability, distilling agent-invariant constraints so one memory can be shared across backbones. uses the same disagreement for a different end: rather than normalizing model differences away, it exploits them to raise memory quality among frontier sources under a matched budget.","intent_p":0.95,"contests_p":null}]},{"kind":"supports_this_paper","label":"Cites as support","papers":7,"sentences":7,"quotes":[{"arxiv_id":"2510.04618","title":"Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models","url":"/paper/2510.04618","section":"Introduction","sentence":"Append-only accumulation compounds this with redundancy and drift: agents over-follow retrieved experience, propagate stale errors, and degrade as continuously updated memories accumulate faults [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2505.16067","title":null,"url":"/paper/2505.16067","section":"Introduction","sentence":"Append-only accumulation compounds this with redundancy and drift: agents over-follow retrieved experience, propagate stale errors, and degrade as continuously updated memories accumulate faults [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2603.10600","title":null,"url":"/paper/2603.10600","section":"Introduction","sentence":"Append-only accumulation compounds this with redundancy and drift: agents over-follow retrieved experience, propagate stale errors, and degrade as continuously updated memories accumulate faults [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2605.12978","title":"Useful Memories Become Faulty When Continuously Updated by LLMs","url":"/paper/2605.12978","section":"Introduction","sentence":"Append-only accumulation compounds this with redundancy and drift: agents over-follow retrieved experience, propagate stale errors, and degrade as continuously updated memories accumulate faults [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2603.23234","title":"MemCollab: Cross-Model Memory Collaboration via Contrastive Trajectory Distillation","url":"/paper/2603.23234","section":"Introduction","sentence":"Examining multiple paths surfaces stronger strategies than any single execution [ref] ; for computer-use agents, such paths need not be consumed at test time and discarded.","intent_p":0.96,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":6,"sentences":7,"quotes":[{"arxiv_id":"2602.11964","title":null,"url":"/paper/2602.11964","section":"Related Work","sentence":"Recent benchmarks stress long-horizon, stateful app-agent behavior—conversational tool use in ToolSandbox, cross-application coding agents in AppWorld, and dynamic asynchronous tasks with ambiguity, time obligations, and action-level write verification in / [ref] —exposing a gap targets.","intent_p":0.8,"contests_p":null},{"arxiv_id":"2603.10600","title":null,"url":"/paper/2603.10600","section":"Related Work","sentence":"Closer to our setting, Trajectory-Informed Memory Generation produces strategy and recovery tips for future AppWorld tasks [ref] , and Contextual Experience Replay keeps a dynamic buffer of environment dynamics for web agents [ref] .","intent_p":0.93,"contests_p":null},{"arxiv_id":"2509.25140","title":"ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory","url":"/paper/2509.25140","section":"Related Work","sentence":"A large body of work converts past executions into reusable experience: verbal reflections (Reflexion), lessons (ExpeL), executable skills (Voyager), induced workflows (Agent Workflow Memory) [ref] , and distilled thought templates or procedural memories [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2305.16291","title":"Voyager: An Open-Ended Embodied Agent with Large Language Models","url":"/paper/2305.16291","section":"Related Work","sentence":"A large body of work converts past executions into reusable experience: verbal reflections (Reflexion), lessons (ExpeL), executable skills (Voyager), induced workflows (Agent Workflow Memory) [ref] , and distilled thought templates or procedural memories [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2409.07429","title":"Agent Workflow Memory","url":"/paper/2409.07429","section":"Related Work","sentence":"A large body of work converts past executions into reusable experience: verbal reflections (Reflexion), lessons (ExpeL), executable skills (Voyager), induced workflows (Agent Workflow Memory) [ref] , and distilled thought templates or procedural memories [ref] .","intent_p":0.97,"contests_p":null}]}],"unlabelled_sentences":10},"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/2608.22533"}}