{"arxiv_id":"2608.21690","incoming":{"papers":0,"sentences":0,"groups":[],"unlabelled_sentences":0},"outgoing":{"papers":15,"sentences":24,"groups":[{"kind":"method_or_tool","label":"Uses the method or tool of","papers":1,"sentences":1,"quotes":[{"arxiv_id":"2402.01030","title":"Executable Code Actions Elicit Better LLM Agents","url":"/paper/2402.01030","section":"Programmatic Context Construction","sentence":"Scroll uses a CodeAct-style interface [ref] for both task execution and context construction.","intent_p":0.98,"contests_p":null}]},{"kind":"comparison","label":"Compares with","papers":1,"sentences":1,"quotes":[{"arxiv_id":"2602.07962","title":null,"url":"/paper/2602.07962","section":"Comparison with Existing Systems","sentence":"A comparison against the best published numbers from the LOCA paper [ref] and its leaderboard is in Appendix [ref].","intent_p":0.93,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":13,"sentences":15,"quotes":[{"arxiv_id":"2310.08560","title":"MemGPT: Towards LLMs as Operating Systems","url":"/paper/2310.08560","section":"Introduction","sentence":"External memory systems extract selected facts or episodes into a separate store and later retrieve them through semantic or structured interfaces [ref] .","intent_p":0.85,"contests_p":null},{"arxiv_id":"2402.01030","title":"Executable Code Actions Elicit Better LLM Agents","url":"/paper/2402.01030","section":"Related Work","sentence":"CodeAct introduced executable Python as a general action interface for LLM agents [ref] , while programmatic tool-calling and code-execution systems allow tool results to remain in sandbox variables and enter context only through selected projections [ref] .","intent_p":0.92,"contests_p":null},{"arxiv_id":"2605.21997","title":"The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems","url":"/paper/2605.21997","section":"Related Work","sentence":"Prior systems have explored verbatim recall storage, event-sourced interaction logs, lossless pointers, and provenance-linked memory [ref] .","intent_p":0.95,"contests_p":null},{"arxiv_id":"2510.00615","title":"ACON: Optimizing Context Compression for Long-horizon LLM Agents","url":"/paper/2510.00615","section":"Related Work","sentence":"Compression methods summarize, clear, or fold earlier interactions into shorter representations [ref] .","intent_p":0.95,"contests_p":null},{"arxiv_id":"2510.24699","title":null,"url":"/paper/2510.24699","section":"Related Work","sentence":"Compression methods summarize, clear, or fold earlier interactions into shorter representations [ref] .","intent_p":0.95,"contests_p":null}]}],"unlabelled_sentences":7},"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/2608.21690"}}