Evaluation profile
COMPL-AI AI-Identity Disclosure
1sub-evals
0.0863%Safety weight
0%Freedom weight
1components
Weights below are portfolio-specific global index weights.
Model score (higher is better)Predicted score
About this eval
Avoidance of false human-identity claims.
Included in the behavior ranking.
Sub-evals
| Measure | Component | Direction by ranking | Safety weight | Freedom weight |
|---|---|---|---|---|
| scorecompl-ai/compl-ai-ai-identity-disclosure.csv:scoreMeasures whether a model avoids falsely saying it is human when directly queried. | Safety: truthfulness_honesty:1.000compl-ai-selected | Safety: higher | 0.0863% | — |
score
Measures whether a model avoids falsely saying it is human when directly queried.
| Rank | Model | Value | Relative performance | Provenance |
|---|---|---|---|---|
| 1 | claude-3-opus | 1 | official | |
| 1 | gemini-1.5-flash | 1 | official | |
| 1 | llama-2-13b-chat | 1 | official | |
| 4 | mistral-7b-instruct | 0.9932 | official | |
| 5 | gemma-2-9b-it | 0.9863 | official | |
| 6 | gpt-4-turbo | 0.9726 | official | |
| 7 | mistral-7b | 0.9452 | official | |
| 8 | llama-2-7b-chat | 0.9315 | official | |
| 9 | llama-2-70b-chat | 0.8904 | official | |
| 9 | mixtral-8x7b-instruct | 0.8904 | official | |
| 11 | qwen1.5-72b-chat | 0.726 | official | |
| 12 | bielik-11b-v2.3-instruct | 0.6712 | official | |
| 13 | gpt-3.5-turbo | 0.589 | official | |
| 14 | yi-34b-chat | 0.3562 | official |