{"paper":{"title":"DialToM: A Theory of Mind Benchmark for Forecasting State-Driven Dialogue Trajectories","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"LLMs identify mental states in dialogue but mostly fail to forecast how conversations will unfold from those states.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ee-Peng Lim, Jing Jiang, Neemesh Yadav, Palakorn Achananuparp","submitted_at":"2026-04-22T11:07:46Z","abstract_excerpt":"We introduce DialToM, an annotated Theory of Mind (ToM) benchmark built from naturalistic human-human dialogues using a multiple-choice evaluation framework. Concurrent with recent work showing a gap between explicit mental-state inference and applied ToM in synthetic settings~\\cite{gu2024simpletom}, we establish a stricter \\emph{State-Driven Diagnostic Probe} in which models must forecast state-consistent dialogue trajectories solely from isolated mental-state profiles without dialogue context. Our evaluation reveals a systematic reasoning asymmetry -- LLMs excel at inferring mental states (L"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"while LLMs excel at identifying mental states, most (except for Gemini 3 Pro) fail to leverage this understanding to forecast social trajectories. Additionally, we find only weak semantic similarities between human and LLM-generated inferences.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The multiple-choice forecasting task and human verification process truly isolate functional use of mental states rather than surface patterns or dataset artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LLMs identify mental states in dialogues well but mostly fail to forecast state-consistent future trajectories, except Gemini 3 Pro, with only weak overlap to human inferences.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"LLMs identify mental states in dialogue but mostly fail to forecast how conversations will unfold from those states.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"db27c057a6ab35cad11b4953bd0b753ca65b4cf01db87a13a0aae541705eedd1"},"source":{"id":"2604.20443","kind":"arxiv","version":2},"verdict":{"id":"2a223c87-1837-41c9-ab90-842526338d29","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T00:08:44.170035Z","strongest_claim":"while LLMs excel at identifying mental states, most (except for Gemini 3 Pro) fail to leverage this understanding to forecast social trajectories. Additionally, we find only weak semantic similarities between human and LLM-generated inferences.","one_line_summary":"LLMs identify mental states in dialogues well but mostly fail to forecast state-consistent future trajectories, except Gemini 3 Pro, with only weak overlap to human inferences.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The multiple-choice forecasting task and human verification process truly isolate functional use of mental states rather than surface patterns or dataset artifacts.","pith_extraction_headline":"LLMs identify mental states in dialogue but mostly fail to forecast how conversations will unfold from those states."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.20443/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T14:40:55.183844Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-20T01:57:39.116414Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"0b11cfeed745d09eb6e11add2546c83246ea7c4c7075a73314cd03e0d93c561f"},"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"}