{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:XRPTGFZFV5L2UKMTEUXDSHSWFO","short_pith_number":"pith:XRPTGFZF","schema_version":"1.0","canonical_sha256":"bc5f331725af57aa2993252e391e562b982564dcaf669610b9c3b279c8dc16bb","source":{"kind":"arxiv","id":"2602.16662","version":2},"attestation_state":"computed","paper":{"title":"Evaluating Collective Behaviour of Hundreds of LLM Agents","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.MA","authors_text":"Jianing Zhao, Joel Z. Leibo, Richard Willis, Yali Du","submitted_at":"2026-02-18T18:02:51Z","abstract_excerpt":"LLM-powered AI assistants acting on behalf of users can produce poor collective outcomes at scale. We introduce a framework for evaluating their emergent behaviour in social dilemmas, applied to three iterated games (Public Goods, Collective Risk, Common Pool Resource). We prompt each model to produce a natural-language strategy, then have the same model translate it into code. This aims to isolate strategic reasoning from input-parsing, enables pre-deployment inspection, and scales to populations of hundreds of agents. We propose three analyses: behavioural fingerprinting via exhaustive evalu"},"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":"2602.16662","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.MA","submitted_at":"2026-02-18T18:02:51Z","cross_cats_sorted":[],"title_canon_sha256":"2f828881b2a9c30de008ac80f7eaea85e8f9c14bac146359d55d41dc5a38f0b8","abstract_canon_sha256":"3419a5c13e2c1ab102d75fcd5f315a0ef1afeac920756e5682d6f4ef7e1c71f2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:54.548645Z","signature_b64":"0u2Ir6apy2cs3SmXyrL3rD35Hj0NML2F8qsr8OGvsQRoPDuG3VQxLkbEUMeLB10+VTyNeh+DLZiV9vz5gz0pDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc5f331725af57aa2993252e391e562b982564dcaf669610b9c3b279c8dc16bb","last_reissued_at":"2026-07-03T01:17:54.548220Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:54.548220Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating Collective Behaviour of Hundreds of LLM Agents","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.MA","authors_text":"Jianing Zhao, Joel Z. Leibo, Richard Willis, Yali Du","submitted_at":"2026-02-18T18:02:51Z","abstract_excerpt":"LLM-powered AI assistants acting on behalf of users can produce poor collective outcomes at scale. We introduce a framework for evaluating their emergent behaviour in social dilemmas, applied to three iterated games (Public Goods, Collective Risk, Common Pool Resource). We prompt each model to produce a natural-language strategy, then have the same model translate it into code. This aims to isolate strategic reasoning from input-parsing, enables pre-deployment inspection, and scales to populations of hundreds of agents. We propose three analyses: behavioural fingerprinting via exhaustive evalu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.16662","kind":"arxiv","version":2},"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/2602.16662/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":"2602.16662","created_at":"2026-07-03T01:17:54.548280+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.16662v2","created_at":"2026-07-03T01:17:54.548280+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.16662","created_at":"2026-07-03T01:17:54.548280+00:00"},{"alias_kind":"pith_short_12","alias_value":"XRPTGFZFV5L2","created_at":"2026-07-03T01:17:54.548280+00:00"},{"alias_kind":"pith_short_16","alias_value":"XRPTGFZFV5L2UKMT","created_at":"2026-07-03T01:17:54.548280+00:00"},{"alias_kind":"pith_short_8","alias_value":"XRPTGFZF","created_at":"2026-07-03T01:17:54.548280+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2606.28456","citing_title":"Is Lying an Emergent Behaviour in LLMs? Evidence from Gaslighting AI agents in a Sustainability Game","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2605.29874","citing_title":"Evolutionary Dynamics of Cooperation in Next-Generation LLM Agent Systems: A Cross-Provider Empirical Extension","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"2604.22452","citing_title":"Superminds Test: Actively Evaluating Collective Intelligence of Agent Society via Probing Agents","ref_index":50,"is_internal_anchor":true},{"citing_arxiv_id":"2604.12162","citing_title":"AlphaEval: Evaluating Agents in Production","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO","json":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO.json","graph_json":"https://pith.science/api/pith-number/XRPTGFZFV5L2UKMTEUXDSHSWFO/graph.json","events_json":"https://pith.science/api/pith-number/XRPTGFZFV5L2UKMTEUXDSHSWFO/events.json","paper":"https://pith.science/paper/XRPTGFZF"},"agent_actions":{"view_html":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO","download_json":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO.json","view_paper":"https://pith.science/paper/XRPTGFZF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.16662&json=true","fetch_graph":"https://pith.science/api/pith-number/XRPTGFZFV5L2UKMTEUXDSHSWFO/graph.json","fetch_events":"https://pith.science/api/pith-number/XRPTGFZFV5L2UKMTEUXDSHSWFO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO/action/storage_attestation","attest_author":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO/action/author_attestation","sign_citation":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO/action/citation_signature","submit_replication":"https://pith.science/pith/XRPTGFZFV5L2UKMTEUXDSHSWFO/action/replication_record"}},"created_at":"2026-07-03T01:17:54.548280+00:00","updated_at":"2026-07-03T01:17:54.548280+00:00"}