{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:NRUDAJ65J47VYHCUTC3KKNZ3DF","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"c5c44621015278e87fe24c7e5eb44ccc9cdd13eafe424e2424c51a0f8874ed69","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-17T18:50:33Z","title_canon_sha256":"c90be175ba0892e0bb9f340f2945ab6ac8db699543ec10a296369e88e1c85c64"},"schema_version":"1.0","source":{"id":"2412.13171","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.13171","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"arxiv_version","alias_value":"2412.13171v1","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.13171","created_at":"2026-05-17T23:38:15Z"},{"alias_kind":"pith_short_12","alias_value":"NRUDAJ65J47V","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"NRUDAJ65J47VYHCU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"NRUDAJ65","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:37665ae78cfc81152dc982748cb2c4a4803ac1f2368109d25735c568b1fd1689","target":"graph","created_at":"2026-05-17T23:38:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Through experiments, we illustrate how CCoT enables additional reasoning over dense contentful representations to achieve corresponding improvements in accuracy. Moreover, the reasoning improvements can be adaptively modified on demand by controlling the number of contemplation tokens generated."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the generated continuous contemplation tokens actually encode and preserve the semantic content of explicit reasoning chains rather than functioning primarily as additional learned parameters or regularizers whose benefit is not tied to interpretable reasoning."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"CCoT generates variable-length continuous contemplation tokens that compress explicit reasoning chains, enabling additional dense reasoning and accuracy gains in off-the-shelf language models while allowing adaptive control of token count."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Language models can reason more accurately by generating compressed continuous tokens that stand in for full reasoning chains."}],"snapshot_sha256":"b938dfd0d3b94388959f5e33794de953f2e4ff4ea69a5650fc8c8898921d4b07"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"44764b041e36faa2c28696dfbc06f0c284413dff4a66788264c6fc50a1c59dfe"},"paper":{"abstract_excerpt":"Chain-of-thought (CoT) decoding enables language models to improve reasoning performance at the cost of high generation latency in decoding. Recent proposals have explored variants of contemplation tokens, a term we introduce that refers to special tokens used during inference to allow for extra computation. Prior work has considered fixed-length sequences drawn from a discrete set of embeddings as contemplation tokens. Here we propose Compressed Chain-of-Thought (CCoT), a framework to generate contentful and continuous contemplation tokens of variable sequence length. The generated contemplat","authors_text":"Benjamin Van Durme, Jeffrey Cheng","cross_cats":[],"headline":"Language models can reason more accurately by generating compressed continuous tokens that stand in for full reasoning chains.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-17T18:50:33Z","title":"Compressed Chain of Thought: Efficient Reasoning Through Dense Representations"},"references":{"count":23,"internal_anchors":10,"resolved_work":23,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"arXiv preprint arXiv:2006.11527 , year=","work_id":"1be92d31-396f-48a6-bc33-0ae9363a096c","year":2006},{"cited_arxiv_id":"2110.14168","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Implicit chain of thought reasoning via knowledge distillation","work_id":"9fdf9899-15b3-4c02-93fb-f7bb42f8f2d6","year":null},{"cited_arxiv_id":"2405.14838","doi":"","is_internal_anchor":true,"ref_index":4,"title":"From Explicit CoT to Implicit CoT: Learning to Internalize CoT Step by Step","work_id":"70541c9d-c3f3-48e4-be00-e7de1622f612","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"In-context autoencoder for context compression in a large language model.arXiv preprint arXiv:2307.06945","work_id":"20a3380e-2b43-4509-a62c-260b1efae1c8","year":null}],"snapshot_sha256":"03758753a25140aeeacf43ab0d5e2d008b16a38c942fefbc4f0daa098498ce85"},"source":{"id":"2412.13171","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-17T04:44:18.665036Z","id":"7b174ffb-b384-4af3-9706-d006f0f547de","model_set":{"reader":"grok-4.3"},"one_line_summary":"CCoT generates variable-length continuous contemplation tokens that compress explicit reasoning chains, enabling additional dense reasoning and accuracy gains in off-the-shelf language models while allowing adaptive control of token count.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Language models can reason more accurately by generating compressed continuous tokens that stand in for full reasoning chains.","strongest_claim":"Through experiments, we illustrate how CCoT enables additional reasoning over dense contentful representations to achieve corresponding improvements in accuracy. Moreover, the reasoning improvements can be adaptively modified on demand by controlling the number of contemplation tokens generated.","weakest_assumption":"That the generated continuous contemplation tokens actually encode and preserve the semantic content of explicit reasoning chains rather than functioning primarily as additional learned parameters or regularizers whose benefit is not tied to interpretable reasoning."}},"verdict_id":"7b174ffb-b384-4af3-9706-d006f0f547de"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3e6a10b85bdbddadafdeb7e7911451cd3555686ba41e9f0cc9c5227437cbf0f0","target":"record","created_at":"2026-05-17T23:38:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"c5c44621015278e87fe24c7e5eb44ccc9cdd13eafe424e2424c51a0f8874ed69","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-17T18:50:33Z","title_canon_sha256":"c90be175ba0892e0bb9f340f2945ab6ac8db699543ec10a296369e88e1c85c64"},"schema_version":"1.0","source":{"id":"2412.13171","kind":"arxiv","version":1}},"canonical_sha256":"6c683027dd4f3f5c1c5498b6a5373b1968ba25b3c53307f3264a913496a2cfdb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6c683027dd4f3f5c1c5498b6a5373b1968ba25b3c53307f3264a913496a2cfdb","first_computed_at":"2026-05-17T23:38:15.084912Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:15.084912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gei5leovLiaIh4VkV0GnpZOqwEaNipbkVeKpXif126DHF0Fj3Y5RWEr/tAkb+23wAExkIEsbZI/m7bRcmLKIDA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:15.085595Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.13171","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3e6a10b85bdbddadafdeb7e7911451cd3555686ba41e9f0cc9c5227437cbf0f0","sha256:37665ae78cfc81152dc982748cb2c4a4803ac1f2368109d25735c568b1fd1689"],"state_sha256":"18f82e1326050556468b03082a5ac1e9c249abf878e26d1f9c070b999bb24d93"}