{"paper":{"title":"From Migration to Calibration: Preserving Agent Capabilities across Models, Jurisdictions, and Scale","arxiv_id":"2609.35149","primary_cat":"cs.AI","submitted_at":"2026-09-28T13:31:46+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":"Agents need calibration when deployment conditions change: replacing a driving model, including a foundation-to-post-trained transition; crossing jurisdictions; or scaling across heterogeneous markets and sources.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.06,"main_claim_p":0.06,"quantitative":false,"quantitative_p":0.09,"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":0.51,"id":"s1","own":false,"role":"background","text":"Interface compatibility alone does not establish capability retention or target-contract satisfaction.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.12,"main_claim_p":0.1,"quantitative":false,"quantitative_p":0.16,"role_probabilities":{"gap":0.48,"method":0.0,"result":0.0,"background":0.51,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.01}},{"i":2,"p":0.75,"id":"s2","own":false,"role":"contribution","text":"We formulate agent calibration as constrained behavioral adaptation across three interacting layers: information preservation, harness adaptation, and user acceptance; the layers apply to every scenario, not one-to-one to the three.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.4,"main_claim_p":0.14,"quantitative":false,"quantitative_p":0.36,"role_probabilities":{"gap":0.0,"method":0.25,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.75,"interpretation":0.0}},{"i":3,"p":0.6,"id":"s3","own":false,"role":"contribution","text":"The basic objective is non-degradation on prespecified capability measures while satisfying target requirements; aggregate improvement is stronger.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.13,"main_claim_p":0.24,"quantitative":true,"quantitative_p":0.74,"role_probabilities":{"gap":0.0,"method":0.27,"result":0.02,"background":0.08,"limitation":0.0,"future_work":0.0,"contribution":0.6,"interpretation":0.03}},{"i":4,"p":0.85,"id":"s4","own":false,"role":"method","text":"Information calibration preserves independently validated source content still applicable to the target task.","hedge":{"p":0.98,"label":"asserted"},"own_p":0.11,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.85,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.15,"interpretation":0.0}},{"i":5,"p":0.86,"id":"s5","own":false,"role":"method","text":"Harness calibration aligns observable artifacts at semantic checkpoints and repairs them through iteration, tool substitution, or local replanning within explicit budgets.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.12,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.19,"role_probabilities":{"gap":0.0,"method":0.86,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.14,"interpretation":0.0}},{"i":6,"p":0.79,"id":"s6","own":false,"role":"method","text":"User calibration enforces recipient-specific output contracts: templates, schemas, and section-level preferences.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.1,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.79,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.21,"interpretation":0.0}},{"i":7,"p":0.47,"id":"s7","own":false,"role":"contribution","text":"A global e-commerce example shows how shared standards coexist with site- and market-specific adapters and validation.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.38,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.08,"role_probabilities":{"gap":0.0,"method":0.47,"result":0.03,"background":0.01,"limitation":0.0,"future_work":0.0,"contribution":0.47,"interpretation":0.02}},{"i":8,"p":0.53,"id":"s8","own":false,"role":"method","text":"We distinguish trainable policies from frozen-backbone configuration or controller optimization, and evidence verification from relative judgment and DPO/GRPO optimization.","hedge":{"p":0.95,"label":"asserted"},"own_p":0.14,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.21,"role_probabilities":{"gap":0.01,"method":0.53,"result":0.0,"background":0.08,"limitation":0.0,"future_work":0.0,"contribution":0.37,"interpretation":0.01}},{"i":9,"p":0.91,"id":"s9","own":false,"role":"background","text":"Recent harness-transfer and judge-validity studies motivate target-native execution records, separate audits of task validity and near-tie ranking, and matched target-native optimization controls.","hedge":{"p":0.79,"label":"asserted"},"own_p":0.06,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.09,"role_probabilities":{"gap":0.01,"method":0.02,"result":0.0,"background":0.91,"limitation":0.0,"future_work":0.0,"contribution":0.06,"interpretation":0.0}},{"i":10,"p":0.47,"id":"s10","own":false,"role":"method","text":"We propose held-out evaluations for model changes, cross-border adaptation, and scale, including a factorial test of source evidence and checkpoint repair and group-level reporting to prevent aggregate gains from masking local failures.","hedge":{"p":0.81,"label":"asserted"},"own_p":0.22,"main_claim_p":0.16,"quantitative":false,"quantitative_p":0.24,"role_probabilities":{"gap":0.0,"method":0.47,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.07,"contribution":0.46,"interpretation":0.0}},{"i":11,"p":0.52,"id":"s11","own":false,"role":"limitation","text":"This is a methodological proposal; implementation and empirical validation remain future work.","hedge":{"p":0.96,"label":"asserted"},"own_p":0.12,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.52,"future_work":0.48,"contribution":0.0,"interpretation":0.0}}]},"compiled":{"main_claim":{"p":0.3,"text":"We formulate agent calibration as constrained behavioral adaptation across three interacting layers: information preservation, harness adaptation, and user acceptance; the layers apply to every scenario, not one-to-one to the three.","location":"abstract s2","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":8771,"cost_usd_reported":0.0},"compiled_at":"2026-09-29","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.05,"states_limitation":0.98,"human_participants":0.11,"controlled_comparison":0.49,"randomized_assignment":0.03,"states_quantitative_result":0.03,"compares_with_prior_results":0.1,"mentions_code_or_data_availability":0.02},"evidential_basis":{"p":0.96,"value":"argument","probabilities":{"argument":0.96,"literature":0.04,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.0,"simulation_or_computation":0.0}},"contribution_type":{"p":0.9,"value":"position or commentary","probabilities":{"review":0.0,"method_or_tool":0.1,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.9}},"contribution_statement":{"p":0.59,"text":"We propose held-out evaluations for model changes, cross-border adaptation, and scale, including a factorial test of source evidence and checkpoint repair and group-level reporting to prevent aggregate gains from masking local failures.","location":"abstract s10"}},"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":12,"sentences_in_abstract":12},"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"}}}