{"paper":{"title":"Remembering Has a Price: Learning Memory Worth under Resource Constraints","arxiv_id":"2610.26453","primary_cat":"cs.AI","submitted_at":"2026-10-19T14:09:45+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Lifelong agents continually accumulate and reuse experience to improve after deployment, yet in practice they often operate under tight generation, interaction, and memory budgets.","hedge":{"p":0.95,"label":"asserted"},"own_p":0.07,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.13,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.82,"id":"s1","own":true,"role":"result","text":"Under such stringent joint constraints, we find that existing lifelong-memory methods can become ineffective, with several sophisticated approaches offering little advantage over simple retrieval or even no-memory agents.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.85,"main_claim_p":0.1,"quantitative":true,"quantitative_p":0.77,"role_probabilities":{"gap":0.08,"method":0.0,"result":0.82,"background":0.1,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.93,"id":"s2","own":true,"role":"interpretation","text":"This reveals a more fundamental problem: resource scarcity makes each memory decision more consequential.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.57,"main_claim_p":0.13,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.04,"method":0.0,"result":0.01,"background":0.02,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.93}},{"i":3,"p":0.47,"id":"s3","own":false,"role":"background","text":"When only a small amount of past experience can be retained or reused and opportunities for reasoning and recovery are limited, retrieving a relevant but ineffective memory can consume scarce context and hinder downstream execution.","hedge":{"p":0.85,"label":"asserted"},"own_p":0.21,"main_claim_p":0.06,"quantitative":false,"quantitative_p":0.22,"role_probabilities":{"gap":0.07,"method":0.0,"result":0.0,"background":0.47,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.46}},{"i":4,"p":0.79,"id":"s4","own":false,"role":"gap","text":"The key challenge is therefore not simply to retrieve relevant experience, but to determine which experiences are actually worth retaining and reusing.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.36,"main_claim_p":0.23,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.79,"method":0.0,"result":0.0,"background":0.18,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.03}},{"i":5,"p":0.57,"id":"s5","own":true,"role":"method","text":"To address this challenge, we propose TokenWorth, a lightweight online memory framework that learns query-dependent memory worth from sparse execution feedback and applies it consistently to memory retrieval and bounded memory management.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.9,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.0,"method":0.57,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.43,"interpretation":0.0}},{"i":6,"p":1.0,"id":"s6","own":true,"role":"result","text":"Across three LifelongAgentBench environments and both local and API-based deployments, TokenWorth achieves the highest task success rate while using the fewest output tokens among the evaluated methods.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.12,"quantitative":true,"quantitative_p":0.96,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":7,"p":1.0,"id":"s7","own":true,"role":"interpretation","text":"These results suggest that effective lifelong adaptation under resource constraints depends not on remembering more, but on learning what is worth remembering.","hedge":{"p":0.99,"label":"hedged_possibility"},"own_p":0.78,"main_claim_p":0.93,"quantitative":false,"quantitative_p":0.24,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":1.0}},{"i":8,"p":0.65,"id":"s8","own":false,"role":"contribution","text":"Our code is publicly available at https://github.com/AIGNLAI/TokenWorth.","hedge":{"p":0.82,"label":"asserted"},"own_p":0.46,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.02,"role_probabilities":{"gap":0.0,"method":0.07,"result":0.01,"background":0.02,"limitation":0.0,"future_work":0.24,"contribution":0.65,"interpretation":0.01}}]},"compiled":{"main_claim":{"p":1.0,"text":"These results suggest that effective lifelong adaptation under resource constraints depends not on remembering more, but on learning what is worth remembering.","location":"abstract s7","hedge_strength":{"p":0.99,"label":"hedged possibility"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":6878,"cost_usd_reported":0.0},"compiled_at":"2026-10-20","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.05,"states_limitation":0.13,"human_participants":0.04,"controlled_comparison":0.89,"randomized_assignment":0.02,"states_quantitative_result":0.21,"compares_with_prior_results":0.97,"mentions_code_or_data_availability":0.98},"evidential_basis":{"p":0.96,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.96,"simulation_or_computation":0.04}},"contribution_type":{"p":1.0,"value":"method or tool","probabilities":{"review":0.0,"method_or_tool":1.0,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":1.0,"text":"To address this challenge, we propose TokenWorth, a lightweight online memory framework that learns query-dependent memory worth from sparse execution feedback and applies it consistently to memory retrieval and bounded memory management.","location":"abstract s5"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":9,"sentences_in_abstract":9},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}