O3's unusual chain-of-thought terms also occur in settings unrelated to deception.
Assessment
The evidence favors the claim, but the examination is incomplete or the evidence is indirect.
OpenAI o3's chain-of-thought contains idiosyncratic vocabulary such as "watchers", "illusions", "overshadow", "vantage", "disclaim" and "marinade", which drew attention because it appears in reasoning about evaluations and covert actions. The Apollo Research and OpenAI anti-scheming study (2025) compared word frequencies in o3's reasoning on its anti-scheming evaluations against o3's reasoning on GPQA, a graduate-level science question set with no deceptive element. The terms appear in the GPQA transcripts too, and on average at a higher rate there (the evaluation-to-GPQA ratio averages about 0.65), which indicates the vocabulary is a general feature of o3's reasoning style rather than a marker specific to deception.
The evidence comes from one study by the model's developer and its evaluation partner, on a single comparison dataset, so it is direct but not independently replicated. It shows the terms are not confined to deception settings; it does not show that they never carry deception-related meaning when they do appear in such settings.
Full reasoning: the evidence and decisions behind this verdict
Primary source: "Stress Testing Deliberative Alignment for Anti-Scheming Training" (Schoen et al., Apollo Research and OpenAI, arXiv 2509.15541; copy at www.medianama.com/wp-content/uploads/2025/09/stress_testing_antischeming.pdf). The paper states that the unusual terms "occur in settings unrelated to our environments or deception, appearing at a strictly lower rate (0.65× on average) in our environments compared to GPQA", with a table giving per-word frequencies relative to an English baseline for GPQA and for o3 in the anti-scheming evaluations. GPQA contains multiple-choice science questions with no oversight or deception framing, so occurrence there directly establishes the claim. The paper's qualitative analysis also notes some terms are polysemantic and used as consistent jargon. Limits: single source authored by the model's developer and partner; frequencies depend on their tokenization and term selection; not independently reproduced since o3's raw chain of thought is not public. No contrary evidence found. Status supported rather than verified because the underlying transcripts cannot be examined directly.
Decomposition
This claim is atomic: it bottoms out in a bedrock fact, an open empirical question, or a value premise, and does not decompose further.
Provenance
Where this claim has been said, linked to its canonical form.
These unusual terms occur in settings unrelated to our environments or deception, appearing at a strictly lower rate (0.65× on average) in our environments compared to GPQA (see Table 6).
Section 4, on whether the unusual vocabulary is specific to deception settings.
Cite this claim: a formal citation with its evidence attached
Contribute
Every judgment on this page is open to challenge. A contribution is evaluated on its merits by the reviewer; if it succeeds the page changes, and if it does not, the reasons are stated. Either way the exchange becomes part of the claim’s public record.
The attention this claim received was paid for by a funded mandate. Funding buys only scheduling: it can make an assessment happen sooner, or reach deeper into a subtree. It has no influence on what the assessment concludes, and none on which claims enter the graph; assessments run under the same public standards whoever pays, funders never see or shape a verdict before anyone else, and mandates that attempt to steer conclusions are refused.
Created by extractor · Oct 30, 2026. Every judgment on this page is accompanied by a reasoning trace.