{"paper":{"title":"Experience-Handoff Reinforcement Learning for Recursive Self-Improvement","arxiv_id":"2611.16409","primary_cat":"cs.LG","submitted_at":"2026-11-12T07:18:30+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":"It is common to train Large Language Models (LLMs) for agentic tasks with Reinforcement Learning (RL) from scalar rewards.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.05,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.05,"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":0.53,"id":"s1","own":false,"role":"background","text":"However, this throws away the fine-grained information about the environment that the agent experienced during the trial (e.g., error logs).","hedge":{"p":1.0,"label":"asserted"},"own_p":0.1,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.44,"method":0.0,"result":0.0,"background":0.53,"limitation":0.03,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.66,"id":"s2","own":false,"role":"gap","text":"To utilize this rich information, recent works wrap the agent in a harness that refines its own inputs through reflection or compaction and hands past experience to future rollouts, but the construction of the handed-off artifacts (e.g., a playbook) is itself not learned.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.04,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.66,"method":0.0,"result":0.0,"background":0.34,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":3,"p":0.55,"id":"s3","own":false,"role":"gap","text":"We want the best of both worlds: gradient updates from a scalar reward, while also utilizing environment information to hand off experience.","hedge":{"p":0.63,"label":"asserted"},"own_p":0.1,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.1,"role_probabilities":{"gap":0.55,"method":0.01,"result":0.0,"background":0.14,"limitation":0.0,"future_work":0.01,"contribution":0.29,"interpretation":0.0}},{"i":4,"p":0.93,"id":"s4","own":true,"role":"method","text":"As a concrete testbed, we fix the Agentic Context Engineering (\\ace{}) harness and train one model to act, reflect, and curate task-local playbooks.","hedge":{"p":0.92,"label":"asserted"},"own_p":0.91,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.14,"role_probabilities":{"gap":0.0,"method":0.93,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.07,"interpretation":0.0}},{"i":5,"p":0.95,"id":"s5","own":true,"role":"method","text":"Our method, \\acerl{}, evaluates candidate experience-handoff artifacts by their downstream effect on future actor rollouts, and co-optimizes the role-specific behaviors (action generator, insight reflector, and playbook curator) under a fixed scaffold.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.94,"main_claim_p":0.1,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":0.95,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.05,"interpretation":0.0}},{"i":6,"p":1.0,"id":"s6","own":true,"role":"result","text":"Across agentic benchmarks, \\acerl{} achieves higher held-out performance than baseline methods.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.95,"main_claim_p":0.16,"quantitative":true,"quantitative_p":0.9,"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":"The results suggest that agent training should optimize not only the actions in the current rollout, but also the model-written experience that future rollouts inherit.","hedge":{"p":1.0,"label":"hedged_possibility"},"own_p":0.79,"main_claim_p":0.9,"quantitative":false,"quantitative_p":0.05,"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}}]},"compiled":{"main_claim":{"p":1.0,"text":"The results suggest that agent training should optimize not only the actions in the current rollout, but also the model-written experience that future rollouts inherit.","location":"abstract s7","hedge_strength":{"p":1.0,"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":6337,"cost_usd_reported":0.0},"compiled_at":"2026-11-13","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.07,"states_limitation":0.32,"human_participants":0.03,"controlled_comparison":0.91,"randomized_assignment":0.02,"states_quantitative_result":0.06,"compares_with_prior_results":0.95,"mentions_code_or_data_availability":0.03},"evidential_basis":{"p":0.92,"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":0.92,"simulation_or_computation":0.08}},"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":"Our method, \\acerl{}, evaluates candidate experience-handoff artifacts by their downstream effect on future actor rollouts, and co-optimizes the role-specific behaviors (action generator, insight reflector, and playbook curator) under a fixed scaffold.","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":8,"sentences_in_abstract":8},"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"}}}