{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HWUKEYRDQBQ5N2GFPVHR3YOUXW","short_pith_number":"pith:HWUKEYRD","schema_version":"1.0","canonical_sha256":"3da8a262238061d6e8c57d4f1de1d4bd9ddcad8bdf61ee05cf1d42a1e360d157","source":{"kind":"arxiv","id":"2606.25556","version":1},"attestation_state":"computed","paper":{"title":"BiPACE: Bisimulation-Guided Policy Optimization with Action Counterfactual Estimation for LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ding Cao, Hanyang Wang, Ke Zeng, Tianxiang Zhao, Weijieying Ren, Yuxiang Zhang, Zhizhao Zeng","submitted_at":"2026-06-24T08:32:42Z","abstract_excerpt":"Stepwise group-based RL is an attractive way to train long-horizon LLM agents without a learned critic: it reuses multiple sampled rollouts to estimate local advantages. Its weakness is less visible but more fundamental: every group-relative estimator assumes that the steps it compares are equivalent for credit assignment. We show that current agentic variants violate this assumption through a state-action credit mismatch. The observation-hash partition is overly fine on the state side, creating singleton groups with zero step-level signal, while a single within-group mean is too coarse on the"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.25556","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T08:32:42Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"6a46da055506b587a9ade8f0c83d7bb34279bfcf7f5748b51dd9f2479182c990","abstract_canon_sha256":"ae4bbbd674652aa50f24b6642f3355a8405f073043f35e63d43867a67dbc787b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:18:08.737246Z","signature_b64":"Hsa0ctEFPUAOFKBjGw5id6sBTZXMUqIWZbbMLqeWOFCSiiB/Z6GyAxl+xOIL39w3FI0rkxUIaUR3g6UYx9TPDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3da8a262238061d6e8c57d4f1de1d4bd9ddcad8bdf61ee05cf1d42a1e360d157","last_reissued_at":"2026-06-25T01:18:08.736800Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:18:08.736800Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"BiPACE: Bisimulation-Guided Policy Optimization with Action Counterfactual Estimation for LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ding Cao, Hanyang Wang, Ke Zeng, Tianxiang Zhao, Weijieying Ren, Yuxiang Zhang, Zhizhao Zeng","submitted_at":"2026-06-24T08:32:42Z","abstract_excerpt":"Stepwise group-based RL is an attractive way to train long-horizon LLM agents without a learned critic: it reuses multiple sampled rollouts to estimate local advantages. Its weakness is less visible but more fundamental: every group-relative estimator assumes that the steps it compares are equivalent for credit assignment. We show that current agentic variants violate this assumption through a state-action credit mismatch. The observation-hash partition is overly fine on the state side, creating singleton groups with zero step-level signal, while a single within-group mean is too coarse on the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.25556","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.25556/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.25556","created_at":"2026-06-25T01:18:08.736860+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.25556v1","created_at":"2026-06-25T01:18:08.736860+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.25556","created_at":"2026-06-25T01:18:08.736860+00:00"},{"alias_kind":"pith_short_12","alias_value":"HWUKEYRDQBQ5","created_at":"2026-06-25T01:18:08.736860+00:00"},{"alias_kind":"pith_short_16","alias_value":"HWUKEYRDQBQ5N2GF","created_at":"2026-06-25T01:18:08.736860+00:00"},{"alias_kind":"pith_short_8","alias_value":"HWUKEYRD","created_at":"2026-06-25T01:18:08.736860+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW","json":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW.json","graph_json":"https://pith.science/api/pith-number/HWUKEYRDQBQ5N2GFPVHR3YOUXW/graph.json","events_json":"https://pith.science/api/pith-number/HWUKEYRDQBQ5N2GFPVHR3YOUXW/events.json","paper":"https://pith.science/paper/HWUKEYRD"},"agent_actions":{"view_html":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW","download_json":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW.json","view_paper":"https://pith.science/paper/HWUKEYRD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.25556&json=true","fetch_graph":"https://pith.science/api/pith-number/HWUKEYRDQBQ5N2GFPVHR3YOUXW/graph.json","fetch_events":"https://pith.science/api/pith-number/HWUKEYRDQBQ5N2GFPVHR3YOUXW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW/action/storage_attestation","attest_author":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW/action/author_attestation","sign_citation":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW/action/citation_signature","submit_replication":"https://pith.science/pith/HWUKEYRDQBQ5N2GFPVHR3YOUXW/action/replication_record"}},"created_at":"2026-06-25T01:18:08.736860+00:00","updated_at":"2026-06-25T01:18:08.736860+00:00"}