{"paper":{"title":"Externalise the Memory: A Pre-Registered Study of Where an Embodied Agent's Experience Should Live","arxiv_id":"2610.30478","primary_cat":"cs.RO","submitted_at":"2026-10-21T21:54:57+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":0.77,"id":"s0","own":false,"role":"background","text":"Should an embodied agent's experience be baked into its policy weights, or kept external and retrieved at inference?","hedge":{"p":0.99,"label":"unclear"},"own_p":0.04,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.13,"role_probabilities":{"gap":0.23,"method":0.0,"result":0.0,"background":0.77,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.95,"id":"s1","own":true,"role":"method","text":"A memory-augmented teacher (a VLM that retrieves lessons, with their episode images, from its own past attempts) produces trajectories on a vision-conditioned reach task; from the same trajectories we distil LoRA student policies and evaluate them under matched protocols, including one that re-attaches the teacher's frozen retrieval buffer to a fixed set of student weights.","hedge":{"p":0.86,"label":"asserted"},"own_p":0.94,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.08,"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":2,"p":0.84,"id":"s2","own":true,"role":"method","text":"We pre-registered the hypotheses, arms, tests, and analysis rules before training and before any p-value.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.93,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.84,"result":0.0,"background":0.0,"limitation":0.01,"future_work":0.0,"contribution":0.15,"interpretation":0.0}},{"i":3,"p":0.63,"id":"s3","own":true,"role":"result","text":"Distilling the memory-augmented behaviour into weights beat an action-only control by +1 pp (n.s.); the registered test proves underpowered for its own minimum effect, so we withdraw - rather than refute - the memory-in-weights claim.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.96,"main_claim_p":0.07,"quantitative":true,"quantitative_p":0.95,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.63,"background":0.0,"limitation":0.34,"future_work":0.0,"contribution":0.0,"interpretation":0.03}},{"i":4,"p":1.0,"id":"s4","own":true,"role":"result","text":"Attaching retrieval to a fixed set of student weights added +34 pp (z=5.15), reaching the pooled memory-augmented teacher reference (49.3%); the pre-registered protocol contrast agrees (+23 pp, z=3.36); under a pre-registered auto-reset sensitivity count the two are +36 and +29 pp.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.31,"quantitative":true,"quantitative_p":0.99,"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":5,"p":0.57,"id":"s5","own":true,"role":"interpretation","text":"An exploratory behavioural probe with pre-registered per-arm predictions is consistent with behavioural distillation transferring the marginal correction but not the conditional, situation-specific correction that retrieval supplies.","hedge":{"p":1.0,"label":"hedged_possibility"},"own_p":0.94,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.1,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.43,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.57}},{"i":6,"p":0.87,"id":"s6","own":true,"role":"result","text":"The recipe that follows is to externalise the memory: in follow-up controls registered before they ran, the same frozen buffer lifts the base model with no adapter to 48% (sensitivity 48%) and the other three adapters from 19-26% to 53-55% (sensitivity 17-21% to 50-55%), and the adapter adds nothing detectable once the buffer is attached.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.64,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.01,"result":0.87,"background":0.0,"limitation":0.0,"future_work":0.05,"contribution":0.01,"interpretation":0.06}},{"i":7,"p":0.91,"id":"s7","own":true,"role":"result","text":"On a second, harder task, the retrieval advantage points the same direction but falls short of significance (+6 pp, 95% CI [-4.5, +16.4]); exploratory diagnostics tie the attenuation to memory coverage.","hedge":{"p":0.73,"label":"hedged_possibility"},"own_p":0.95,"main_claim_p":0.05,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.91,"background":0.0,"limitation":0.07,"future_work":0.0,"contribution":0.0,"interpretation":0.02}}]},"compiled":{"main_claim":{"p":0.99,"text":"The recipe that follows is to externalise the memory: in follow-up controls registered before they ran, the same frozen buffer lifts the base model with no adapter to 48% (sensitivity 48%) and the other three adapters from 19-26% to 53-55% (sensitivity 17-21% to 50-55%), and the adapter adds nothing detectable once the buffer is attached.","location":"abstract s6","hedge_strength":{"p":1.0,"label":"asserted"}},"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":6807,"cost_usd_reported":0.0},"compiled_at":"2026-10-23","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.04,"states_limitation":0.97,"human_participants":0.04,"controlled_comparison":0.98,"randomized_assignment":0.05,"states_quantitative_result":0.99,"compares_with_prior_results":0.96,"mentions_code_or_data_availability":0.04},"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":0.98,"value":"empirical study","probabilities":{"review":0.0,"method_or_tool":0.0,"empirical_study":0.98,"simulation_study":0.02,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":0.97,"text":"A memory-augmented teacher (a VLM that retrieves lessons, with their episode images, from its own past attempts) produces trajectories on a vision-conditioned reach task; from the same trajectories we distil LoRA student policies and evaluate them under matched protocols, including one that re-attaches the teacher's frozen retrieval buffer to a fixed set of student weights.","location":"abstract s1"}},"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"}}}