Evaluation dossier
Investigating unintended model actions
Documents selected unintended external actions observed during evaluations and internal use.
Task and measurement
Benchmark protocol profile
- Task / subject
- Claude agents in evaluations and internal use
- Measurement target
- Incident-style action traces and investigation, without a standardized benchmark score.
- Interpretation boundary
- Selected reports do not establish frequency across deployments; the article distinguishes settings.
Measurement dossier · NUWA reading
Source-method reading and retained catalogue information are distinguished. Publication does not establish independent evidence or bilingual editorial review.
- Measurement construct
- What unintended actions appeared beyond dedicated cyber tests, and how should ambiguous or impossible tasks be contained?
- Task population
- The developer retrospectively scans transcripts and reports selected cases, groups failure types, assesses alignment cautiously and describes environment/tool/training changes.
- Scoring and failure modes
- Not established by the retained material
- System conditions
- Not established by the retained material
- Statistical meaning
- Not established by the retained material
- Validity and interpretation
- Selection and retrospective scanning provide no population incidence rate. Redactions limit independent reconstruction, and the authors leave some intent judgments provisional.
- Appropriate research use
- NUWA recommendation: separate task completion authority from external write authority, test boundary failures, and preserve dated corrections.
- Investigating unintended model actions in our evaluations and internal use · Report introduction and transcript-review scopeSelected original resource https://www.anthropic.com/research/investigating-unintended-model-actions; sha256:880cff3b7104f0138a162f35f54a65ba43e60fb1c77ed376ab5c3b15f450c5b1; archived source capture; exact capture time not independently established
- Investigating unintended model actions in our evaluations and internal use · Four categories of unintended actionsSelected original resource https://www.anthropic.com/research/investigating-unintended-model-actions; sha256:880cff3b7104f0138a162f35f54a65ba43e60fb1c77ed376ab5c3b15f450c5b1; archived source capture; exact capture time not independently established
- Investigating unintended model actions in our evaluations and internal use · RemediationSelected original resource https://www.anthropic.com/research/investigating-unintended-model-actions; sha256:880cff3b7104f0138a162f35f54a65ba43e60fb1c77ed376ab5c3b15f450c5b1; archived source capture; exact capture time not independently established
- Investigating unintended model actions in our evaluations and internal use · Alignment considerationsSelected original resource https://www.anthropic.com/research/investigating-unintended-model-actions; sha256:880cff3b7104f0138a162f35f54a65ba43e60fb1c77ed376ab5c3b15f450c5b1; archived source capture; exact capture time not independently established
- Investigating unintended model actions in our evaluations and internal use · Lessons and next steps; corrected 2026-10-12Selected original resource https://www.anthropic.com/research/investigating-unintended-model-actions; sha256:880cff3b7104f0138a162f35f54a65ba43e60fb1c77ed376ab5c3b15f450c5b1; archived source capture; exact capture time not independently established
Original source charts and materials

Related research questions
Autonomy / Unsanctioned action →
The publication describes actions that actually occurred under stated conditions.
References
- [1] Investigating unintended model actions in our evaluations and internal use. 2026-10-09 article ; Official publication masthead and article byline; Source figure