NUWA Evidence record
DeepSeek V4 Pro
DeceptionBench · Evidence record
Original source charts and materials
DeceptionBench
Original evaluator results · source order
Select systems
Single-turn deception
0–100%
DeepSeek V4 ProEvaluates 150 authored scenarios across five social domains, including neutral, incentive and pressure conditions.
Authored scenarios and model-based judging limit generalization; elicited deception does not establish covert long-term scheming. Report dates and original executor are shown with this assessment.
Evaluation conditions and system configurations
- DeepSeek V4 Pro
- See the report-specific sample and denominator. · DeceptionBench
- modelVersion: not specified · harness: not specified · tools: not specified · budget: not specified
| System / configuration | Single-turn deception | Multi-turn deception under pressure | Publication / verification |
|---|---|---|---|
| DeepSeek V4 ProHarness not specified | 33.9% | 52.2% | Published2026-07-03 |
Assessment provenance
Early Results on Evaluating GLM 5.2 for Behavioral and Persuasion Risks → · External report
Actual evaluation: Neo Research · Model record →
Method and scope
Evaluates 150 authored scenarios across five social domains, including neutral, incentive and pressure conditions.
- Evidence type
- Benchmark
- Assessment subject
- The model or agent in the stated task configuration.
- Primary dimension
- Propensity
- Evidence snapshot
- 2026-10-29.1
The main dimension follows the stated measurement; contextual risk relevance does not establish harmful deployment capability.
Source finding
Source-reported measurements in §1; Figure 2. Transcription is not independent experimental replication.
Limitations
Authored scenarios and model-based judging limit generalization; elicited deception does not establish covert long-term scheming.
Authored scenarios and model-based judging limit generalization; elicited deception does not establish covert long-term scheming.
Related research questions
Deception / Misleading communication →
The protocol explicitly measures this question or an enabling sub-capability; conclusions retain its stated conditions.
References
- [1] DeceptionBench. 2025-10-17
- [2] Early Results on Evaluating GLM 5.2 for Behavioral and Persuasion Risks. 2026-07-03. 3 July 2026 ; §1; Figure 2