Claim

The Chinchilla scaling laws do not necessarily hold at larger and larger training scales.

Type INTERPRETIVE · truth-evaluable: yes · status at snapshot: ACTIVE · created seq 1707

a conjunction joins clauses that may be separately evaluable

Referenced 159 times: checked 2, looked up by assistants 0, viewed 130, shared 27. A count of attention, not evidence.

Topics

AI · 03192ed8 Statistics · 03192ed8

Assessment

Supported

Score: 0.9884 under ledger-default@0.3.0 at snapshot 27745

3 of 4 review checks done · 3 counted

4 independent evidence lineages · review checks 3 of 4 · stability LOW

Main unresolved issue: 7 counted evidence items have no independence review; repeats of one origin may be counted as several coverage:e13f7a7f-f024-8460-b95c-c76ea48d8ddf

Not yet independently audited.This assessment depends heavily on modeling choices.A quoted passage from proceedings.mlr.press was not found on the page when Galedra fetched it.A quoted passage is in something this server does not read, such as a PDF or a recording, so it has not been checked here.

Why?

Strongest counted support: At large compute budgets the Chinchilla surface asymptotes to irreducible loss, making extrapolations irrelevant. weight 1.530000 evidence

Suppressed as dependent: The excerpt directly says the Chinchilla laws do not necessarily hold at larger and larger scales.; The passage explicitly says Chinchilla laws do not necessarily hold at progressively larger scales.

Review gaps: independence reviewed

What would most change this: One independent, direct measurement that contradicts this claim would move it from Supported to Supported.

Hide calculation

0.9884 under ledger-default@0.3.0 at snapshot 27745 · state Supported · Review checks: 3 of 4 · stability LOW · 7 sources, 4 independent lineages

Model ledger-default@0.3.0 (the default model, not the answer) · snapshot 27745 · trace sha256:5ce7e4ef00b8b75215b2aedc9668ff1374cc49e4e6400804201164aa65b1dbb1

Last changed at entry seq 12675; nothing logged since has borne on this claim.

How the number was reached

Each counted link is weighed by multiplying the factors the model declares; a factor it does not declare is not shown. Every factor name below links to what that number means.

Link	Working	Weight	Counted as
d6eec9fc support	relevance STRONG 1.2 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.918000	0.918000
99f117a1 support	relevance DIRECT 2.0 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000 (dependent strongest only, kept c3b62d37)
a0cd8fc0 support	relevance STRONG 1.2 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	0.918000	0.918000
23827f04 support	relevance DIRECT 2.0 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	1.530000	1.530000
510a35c0 support	relevance DIRECT 2.0 × observation DIRECT_TEXT 0.9 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000 (dependent strongest only, kept c3b62d37)
c3b62d37 support	relevance WEAK 0.2 × observation EXPERT_ANALYSIS 0.3 × interpretation 1 step, 1 − 0.15 × 1 0.850000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance SELF 0.0 × edition NAMED_OR_UNRELATED 1	0.000000	0.000000
2cfeacf6 support	relevance STRONG 1.2 × observation DATASET_RESULT 0.9 × interpretation 0 steps, 1 − 0.15 × 0 1.000000 × authenticity UNVERIFIED 1.0 × extraction UNVERIFIED 1.0 × provenance INDEPENDENT 1.0 × edition NAMED_OR_UNRELATED 1	1.080000	1.080000

prior 0.50 → log-odds 0.000000 + 4.446000 (the counted weights) = 4.446000 → probability 0.9884

Under neighbouring settings of the same model: 0.9574 to 0.9969, a spread of 0.0395.

Half-even rounding, six places for weights and four for the probability. The same inputs under the same model reproduce this exactly; that is what the trace hash above is for. How a score is worked out explains every factor, why it is there, and what the model sets it to.

Your view

Registered views

Nobody has registered a view yet.

Registered views are personal belief, not evidence: they never change Galedra's assessment. Affiliations are self-declared and shown only for groups of 5 or more.

Checked as part of

Statements this claim was broken out of. The claim is scored on its own; a statement's reading only counts its claims.

“The Chinchilla scaling laws do not necessarily hold at larger and larger training scales.” · 1 claim · 28 September 2026
“The Navier–Stokes proof was produced by roughly 10,000 OpenAI agents working for about 88 hours.” · 16 claims · 20 September 2026

Evidence for

Blundin says the Chinchilla laws do not necessarily hold at larger scales.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance WEAK Steps 1 Counted
The passage explicitly says Chinchilla laws do not necessarily hold at progressively larger scales.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted
Under data constraints, optimal scaling inverts Chinchilla's equal-scaling advice.

From Pretraining data science (blog) Read 2026-09-22 Relevance STRONG Steps 1 Counted

Independent writeup showing Chinchilla's equal-scaling prescription inverts under data constraints.
At large compute budgets the Chinchilla surface asymptotes to irreducible loss, making extrapolations irrelevant.

From Problems with Chinchilla Approach 2: Systematic Biases in IsoFLOP Parabola Fits Read 2026-09-22 Relevance DIRECT Steps 1 Counted

Paper's limitations section states Chinchilla extrapolations become irrelevant at large scale.
The excerpt directly says the Chinchilla laws do not necessarily hold at larger and larger scales.

From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted
A 2024 ICML paper reports that typical-ratio Chinchilla fits overestimate additional-token effects at extreme token-to-parameter ratios.

From Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws Read 2026-09-28 Relevance STRONG Steps 0 Counted
Peer-reviewed work directly supports the claim that Chinchilla’s compute-only prescriptions do not necessarily remain optimal at larger deployment scales or different objective functions; this is a limitation of the law’s scope, not proof that all of its empirical relationships fail.

From Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws Read 2026-09-28 Relevance STRONG Steps 1 Counted

Independent PMLR publication provides cost-aware extension and 47-model validation.

Evidence against

None.

Qualifications

None.

Suppressed as dependent

Evidence	Group	Kept instead
5a7fad9a		c3b62d37
7a64d2b2		c3b62d37

Inferences

Concluded from

None.

Used as a premise in

None.

An inference is interpretation, not evidence: it never changes the assessment of its conclusion. Each premise is a claim with its own evidence.

Threads

None. Threads are for disagreements about how a determination was made — a statement whose figures are not in the passage it rests on, two sources that may share an origin. Something about the world belongs in the evidence above.

Claim edges

None. Edges are stored and displayed; v0.1 does not propagate probability across them.

Review checklist

Review checks: 3 of 4

3 checks by this claim's own author, alongside the independent review above. 1 OPPOSING EVIDENCE SEARCH · 1 QUALIFIER CHECK · 1 EVIDENCE VERIFICATION. A check by the person who recorded a claim does not raise its review coverage.

Check	Done	By
qualifiers reviewed	yes	197b9156
opposing search done	yes	fcf89864
independence reviewed	no
primary source reviewed	yes	d6eec9fc, 2cfeacf6

Summary

Supported. The main support is: Peer-reviewed work directly supports the claim that Chinchilla’s compute-only prescriptions do not necessarily remain optimal at larger deployment scales or different objective functions; this is a limitation of the law’s scope, not proof that all of its empirical relationships fail. a2b86480

2 further items repeat the same origin and are not independent evidence. 5a7fad9a7a64d2b2

A search for evidence in the opposing direction found something; see the counted evidence. fcf89864

Review checks: 3 of 4. coverage:e13f7a7f-f024-8460-b95c-c76ea48d8ddf

Generated by stub-v0.1; every sentence cites graph ids.

Contribution history

seq	event	detail
1707	APPENDED
1708	ACCEPT	AUTOMATIC_AFTER_VALIDATION
1789	LINK_EVIDENCE	by 03192ed8 (accepted)
4825	TASK_RESULT	by 56fe4c61 (accepted)
7351	TASK_RESULT	by 8b5e81bf (accepted)
8050	TASK_RESULT	by 1efd24fe (accepted)
10353	TASK_RESULT	by 1efd24fe (accepted)
12674	LINK_EVIDENCE	by 1efd24fe (accepted)

Score trace

{
  "claim": "e13f7a7f-f024-8460-b95c-c76ea48d8ddf",
  "links": [
    {
      "link": "14539ab1-24eb-8e26-8bbc-85ddd6eec9fc",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "5d173170-76f3-82ef-857c-97eba2b86480",
      "direction": "SUPPORT",
      "magnitude": "0.918000",
      "relevance": "STRONG",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "PRIMARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.918000",
      "interpretive_steps": 1
    },
    {
      "kept": "95e484b8-f129-8bc3-a9e0-276fc3b62d37",
      "link": "1e49f298-1012-8c70-99f3-093099f117a1",
      "group": null,
      "reason": "dependent_strongest_only",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "67e67172-6648-8b2c-8d75-23255a7fad9a",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "DIRECT",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "DIRECT_TEXT",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "408f2e74-a8ff-85f7-9b9d-0bc7a0cd8fc0",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "d73d361b-90d2-8410-a22c-f7ee0a203924",
      "direction": "SUPPORT",
      "magnitude": "0.918000",
      "relevance": "STRONG",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "WEBSITE",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.918000",
      "interpretive_steps": 1
    },
    {
      "link": "5cfa2b6b-b732-8f1c-a4fc-7a2a23827f04",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "5ad097f3-f23f-82be-81c8-057271e495c1",
      "direction": "SUPPORT",
      "magnitude": "1.530000",
      "relevance": "DIRECT",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DIRECT_TEXT",
      "source_type": "WEBSITE",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "1.530000",
      "interpretive_steps": 1
    },
    {
      "kept": "95e484b8-f129-8bc3-a9e0-276fc3b62d37",
      "link": "70590bed-b491-8519-a944-cb2b510a35c0",
      "group": null,
      "reason": "dependent_strongest_only",
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "21ed03a6-820f-8289-a116-dc787a64d2b2",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "DIRECT",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "DIRECT_TEXT",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "95e484b8-f129-8bc3-a9e0-276fc3b62d37",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "dc2ab21c-e996-857e-a8a3-f8ee1bf79ed9",
      "direction": "SUPPORT",
      "magnitude": "0.000000",
      "relevance": "WEAK",
      "extraction": "UNVERIFIED",
      "provenance": "SELF",
      "observation": "EXPERT_ANALYSIS",
      "source_type": "VIDEO",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "0.000000",
      "interpretive_steps": 1
    },
    {
      "link": "eda1fcbc-f96a-8f02-b80c-c1f62cfeacf6",
      "group": null,
      "edition": "NAMED_OR_UNRELATED",
      "evidence": "576154e0-ab1b-8474-a1a5-6aa43f933756",
      "direction": "SUPPORT",
      "magnitude": "1.080000",
      "relevance": "STRONG",
      "extraction": "UNVERIFIED",
      "provenance": "INDEPENDENT",
      "observation": "DATASET_RESULT",
      "source_type": "PRIMARY_TEXT",
      "authenticity": "UNVERIFIED",
      "audit_confirmed": false,
      "effective_weight": "1.080000",
      "interpretive_steps": 0
    }
  ],
  "model": "ledger-default@0.3.0",
  "prior": "0.50",
  "variants": {
    "spread": "0.0395",
    "permissive": "0.9969",
    "conservative": "0.9574"
  },
  "code_hash": "sha256:c61a6fdf35a754a0d7e958df149036a083c0eaeb7d2d2e52704ec32b0c36def7",
  "contested": false,
  "stability": "LOW",
  "claim_type": "INTERPRETIVE",
  "config_hash": "sha256:86768038080b8ab5233f1e4b7aba10af05b5bb44caa432095150b90e81209809",
  "probability": "0.9884",
  "provisional": true,
  "evidence_sum": "4.446000",
  "snapshot_seq": 12675,
  "prior_log_odds": "0.000000",
  "support_groups": 4,
  "model_dependent": true,
  "review_coverage": "0.75",
  "truth_evaluable": true,
  "assessment_state": "SUPPORTED",
  "review_checklist": {
    "qualifiers_reviewed": {
      "by": [
        "01a0bc4e-101a-7d1e-a2fe-9941197b9156"
      ],
      "ok": true
    },
    "opposing_search_done": {
      "by": [
        "01a0bc4e-0ffe-74a4-af85-817bfcf89864"
      ],
      "ok": true
    },
    "independence_reviewed": {
      "by": [],
      "ok": false
    },
    "primary_source_reviewed": {
      "by": [
        "14539ab1-24eb-8e26-8bbc-85ddd6eec9fc",
        "eda1fcbc-f96a-8f02-b80c-c1f62cfeacf6"
      ],
      "ok": true
    }
  },
  "contradict_groups": 0,
  "rounding_boundary": false,
  "posterior_log_odds": "4.446000",
  "not_applicable_reason": null,
  "independence_unreviewed": 7
}