Contribution seq 10353

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "pmlr_beyond_chinchilla_2024",
      "title": "Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws",
      "publisher": "Proceedings of Machine Learning Research",
      "source_type": "PRIMARY_TEXT",
      "retrieved_at": "2026-09-28T21:55:00Z",
      "canonical_uri": "https://proceedings.mlr.press/v235/sardana24a.html",
      "publication_date": "2024-07-01"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "pmlr_chinchilla_excerpt",
      "excerpt": "We train 47 models ... and find that model quality continues to improve as we scale tokens per parameter to extreme ranges (up to 10,000). Finally ... developing scaling laws only from data collected at typical token/parameter ratios overestimates the impact of additional tokens at these extreme ranges.",
      "locator": {},
      "source_id": "pmlr_beyond_chinchilla_2024",
      "excerpt_hash": "sha256:fd583270468479a78f5a2918ae11b43876046062858f376014dc834117f89e75",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "pmlr_chinchilla_evidence",
      "statement": "A 2024 ICML paper reports that typical-ratio Chinchilla fits overestimate additional-token effects at extreme token-to-parameter ratios.",
      "observation_type": "DATASET_RESULT",
      "source_location_id": "pmlr_chinchilla_excerpt"
    },
    {
      "op": "LINK_EVIDENCE",
      "claim_id": "e13f7a7f-f024-8460-b95c-c76ea48d8ddf",
      "direction": "SUPPORT",
      "evidence_item_id": "pmlr_chinchilla_evidence",
      "interpretive_steps": 0,
      "relevance_strength": "STRONG"
    }
  ],
  "outcome": "FOUND"
}

Rows created

• Source 084c0ab4-2fc6-85ec-af33-3ac5178164d1
• SourceLocation af7e8b22-c4ad-89e7-bef7-3ba56f687070
• EvidenceItem 576154e0-ab1b-8474-a1a5-6aa43f933756
• EvidenceClaimLink eda1fcbc-f96a-8f02-b80c-c1f62cfeacf6 (accepted seq 10354)

Status history

Audits

None.

Audit sampling (audit-policy-v0.1): probability 1.0000, sampled yes; inputs {"n":"0.00","mean":"0.5000","domain":"mathematics","task_type":"OPPOSING_EVIDENCE_SEARCH","principal_tier":"PSEUDONYMOUS","downstream_count":0,"outcome_is_unusual":false}