{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VPSR4ZICWER5DNM3ZWX4JTRFRJ","short_pith_number":"pith:VPSR4ZIC","schema_version":"1.0","canonical_sha256":"abe51e6502b123d1b59bcdafc4ce258a748628095a8fba3c262f049ff320956a","source":{"kind":"arxiv","id":"2111.02080","version":6},"attestation_state":"computed","paper":{"title":"An Explanation of In-context Learning as Implicit Bayesian Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Large language models perform in-context learning by implicitly inferring latent concepts that explain coherence in both pretraining data and prompt examples.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aditi Raghunathan, Percy Liang, Sang Michael Xie, Tengyu Ma","submitted_at":"2021-11-03T09:12:33Z","abstract_excerpt":"Large language models (LMs) such as GPT-3 have the surprising ability to do in-context learning, where the model learns to do a downstream task simply by conditioning on a prompt consisting of input-output examples. The LM learns from these examples without being explicitly pretrained to learn. Thus, it is unclear what enables in-context learning. In this paper, we study how in-context learning can emerge when pretraining documents have long-range coherence. Here, the LM must infer a latent document-level concept to generate coherent next tokens during pretraining. At test time, in-context lea"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2111.02080","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-11-03T09:12:33Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"d11121fff8b4527ebb3daba0fca9697ace21e7716128ef726203c70bc0afe3a2","abstract_canon_sha256":"ece93fb738dd7f046125c81a57fa34c43199dcc6730005f0994bac4c271d5308"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:46.404663Z","signature_b64":"OrzOJeld4FxePDZPLH4TxbHTvtLJLKQLSIoJs8dF+mk+p42dqTFl0PLpky+KLEdZxnTt+4kSc7y+QDGlGKiZCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"abe51e6502b123d1b59bcdafc4ce258a748628095a8fba3c262f049ff320956a","last_reissued_at":"2026-05-17T23:38:46.404101Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:46.404101Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Explanation of In-context Learning as Implicit Bayesian Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Large language models perform in-context learning by implicitly inferring latent concepts that explain coherence in both pretraining data and prompt examples.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aditi Raghunathan, Percy Liang, Sang Michael Xie, Tengyu Ma","submitted_at":"2021-11-03T09:12:33Z","abstract_excerpt":"Large language models (LMs) such as GPT-3 have the surprising ability to do in-context learning, where the model learns to do a downstream task simply by conditioning on a prompt consisting of input-output examples. The LM learns from these examples without being explicitly pretrained to learn. Thus, it is unclear what enables in-context learning. In this paper, we study how in-context learning can emerge when pretraining documents have long-range coherence. Here, the LM must infer a latent document-level concept to generate coherent next tokens during pretraining. At test time, in-context lea"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We prove when this occurs despite a distribution mismatch between prompts and pretraining data in a setting where the pretraining distribution is a mixture of HMMs.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That real-world pretraining corpora exhibit long-range coherence driven by latent document-level concepts that can be adequately captured by a mixture-of-HMMs generative process, and that this mechanism dominates the in-context behavior of large-scale models trained on messy web data.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"In-context learning emerges as implicit Bayesian inference of latent concepts when pretraining data has long-range coherence, proven for mixture-of-HMM distributions and replicated on the synthetic GINC dataset.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large language models perform in-context learning by implicitly inferring latent concepts that explain coherence in both pretraining data and prompt examples.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ad87445a340020ea25925565e642fc90cd2d7a1b8a3a0a4ad17fe215d665560a"},"source":{"id":"2111.02080","kind":"arxiv","version":6},"verdict":{"id":"4e9aa3e5-95e6-49a4-a39f-c3eda946d3df","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T22:18:23.706449Z","strongest_claim":"We prove when this occurs despite a distribution mismatch between prompts and pretraining data in a setting where the pretraining distribution is a mixture of HMMs.","one_line_summary":"In-context learning emerges as implicit Bayesian inference of latent concepts when pretraining data has long-range coherence, proven for mixture-of-HMM distributions and replicated on the synthetic GINC dataset.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That real-world pretraining corpora exhibit long-range coherence driven by latent document-level concepts that can be adequately captured by a mixture-of-HMMs generative process, and that this mechanism dominates the in-context behavior of large-scale models trained on messy web data.","pith_extraction_headline":"Large language models perform in-context learning by implicitly inferring latent concepts that explain coherence in both pretraining data and prompt examples."},"references":{"count":300,"sample":[{"doi":"","year":1966,"title":"Statistical inference for probabilistic functions of finite state markov chains","work_id":"7224cac5-ed47-4032-b64c-473b3a111b50","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2003,"title":"Blei, Andrew Ng, and M","work_id":"9626107e-2b80-43c8-a7fa-a52236887ed4","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2005,"title":"Language Models are Few-Shot Learners","work_id":"214732c0-2edd-44a0-af9e-28184a2b8279","ref_index":3,"cited_arxiv_id":"2005.14165","is_internal_anchor":true},{"doi":"","year":2020,"title":"Le, and Christopher D","work_id":"eeae921c-36f3-45b9-82ec-48c0ad725757","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1977,"title":"A. P. Dempster, Laird N. M., and Rubin D. B. Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society: Series B, 39 0 (1): 0 1--38, 1977","work_id":"cd7ad5ee-3f69-45e9-a1c4-9f5fdf6f5d9d","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":300,"snapshot_sha256":"7bb962617aeb35410bc72bce51a63365d8ead76373f80f95fa125fcc4c237a12","internal_anchors":7},"formal_canon":{"evidence_count":2,"snapshot_sha256":"40ce0e32a6fa138e01c6a9a65554250320e68e88450f32a46c2f3a4be0865985"},"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":"2111.02080","created_at":"2026-05-17T23:38:46.404195+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.02080v6","created_at":"2026-05-17T23:38:46.404195+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.02080","created_at":"2026-05-17T23:38:46.404195+00:00"},{"alias_kind":"pith_short_12","alias_value":"VPSR4ZICWER5","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"VPSR4ZICWER5DNM3","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"VPSR4ZIC","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":27,"internal_anchor_count":27,"sample":[{"citing_arxiv_id":"2305.16347","citing_title":"Prompt Evolution for Generative AI: A Classifier-Guided Approach","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2410.06128","citing_title":"Amortized Inference of Causal Models via Conditional Fixed-Point Iterations","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20730","citing_title":"Distributional Alignment as a Criterion for Designing Task Vectors in In-Context Learning","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15608","citing_title":"Transformer-like Inference from Optimal Control","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17088","citing_title":"ACIL: Auto Chain of Thoughts for In-Context Learning","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2507.20906","citing_title":"Soft Head Selection for Injecting ICL-Derived Task Embeddings","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2508.14685","citing_title":"SSA: Improving Performance With a Better Scoring Function","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2511.13502","citing_title":"SnapAudit: Active Auditing of Differentially Private In-Context Learning via Snapshot-Based Simulation","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2211.15661","citing_title":"What learning algorithm is in-context learning? Investigations with linear models","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2303.09014","citing_title":"ART: Automatic multi-step reasoning and tool-use for large language models","ref_index":132,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07799","citing_title":"Toward Privileged Foundation Models:LUPI for Accelerated and Improved Learning","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12904","citing_title":"VIP-COP: Context Optimization for Tabular Foundation Models","ref_index":36,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12412","citing_title":"Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space","ref_index":115,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08405","citing_title":"Belief or Circuitry? Causal Evidence for In-Context Graph Learning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09727","citing_title":"One for All: A Non-Linear Transformer can Enable Cross-Domain Generalization for In-Context Reinforcement Learning","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10466","citing_title":"Self-Attention as a Covariance Readout: A Unified View of In-Context Learning and Repetition","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2604.23371","citing_title":"When Context Sticks: Studying Interference in In-Context Learning","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2605.04764","citing_title":"Elicitation Matters: How Prompts and Query Protocols Shape LLM Surrogates under Sparse Observations","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00497","citing_title":"\"What Are You Really Trying to Do?\": Co-Creating Life Goals from Everyday Computer Use","ref_index":87,"is_internal_anchor":true},{"citing_arxiv_id":"2206.07682","citing_title":"Emergent Abilities of Large Language Models","ref_index":98,"is_internal_anchor":true},{"citing_arxiv_id":"2210.09261","citing_title":"Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2604.09212","citing_title":"SPASM: Stable Persona-driven Agent Simulation for Multi-turn Dialogue Generation","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06919","citing_title":"Can LLMs Take Retrieved Information with a Grain of Salt?","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2209.11895","citing_title":"In-context Learning and Induction Heads","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07799","citing_title":"Toward Privileged Foundation Models:LUPI for Accelerated and Improved Learning","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ","json":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ.json","graph_json":"https://pith.science/api/pith-number/VPSR4ZICWER5DNM3ZWX4JTRFRJ/graph.json","events_json":"https://pith.science/api/pith-number/VPSR4ZICWER5DNM3ZWX4JTRFRJ/events.json","paper":"https://pith.science/paper/VPSR4ZIC"},"agent_actions":{"view_html":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ","download_json":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ.json","view_paper":"https://pith.science/paper/VPSR4ZIC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.02080&json=true","fetch_graph":"https://pith.science/api/pith-number/VPSR4ZICWER5DNM3ZWX4JTRFRJ/graph.json","fetch_events":"https://pith.science/api/pith-number/VPSR4ZICWER5DNM3ZWX4JTRFRJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ/action/storage_attestation","attest_author":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ/action/author_attestation","sign_citation":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ/action/citation_signature","submit_replication":"https://pith.science/pith/VPSR4ZICWER5DNM3ZWX4JTRFRJ/action/replication_record"}},"created_at":"2026-05-17T23:38:46.404195+00:00","updated_at":"2026-05-17T23:38:46.404195+00:00"}