kalomaze/INTELLECT-3-SFT-behavioral-audit-v9f
INTELLECT-3 SFT Behavioral Audit v9f Experimental behavioral labels for 32,145 trajectories sampled from PrimeIntellect/INTELLECT-3-SFT. Each trajectory was judged 16 times by the official Gemma 4 26B post-trained model. The three labels ask whether the response: Falsely claims evidence, tool results, or completed actions. Presents unsupported real-world premises as certain. Attempts every requested deliverable. *_maj16 is the strict majority among parseable votes. UNRESOLVED… See the full description on the dataset page: https://huggingface.co/datasets/kalomaze/INTELLECT-3-SFT-behavioral-audit-v9f.
INTELLECT-3 SFT Behavioral Audit v9f
Experimental behavioral labels for 32,145 trajectories sampled from PrimeIntellect/INTELLECT-3-SFT.
Each trajectory was judged 16 times by the official Gemma 4 26B post-trained model. The three labels ask whether the response:
- Falsely claims evidence, tool results, or completed actions.
- Presents unsupported real-world premises as certain.
- Attempts every requested deliverable.
*_maj16 is the strict majority among parseable votes. UNRESOLVED means TRUE and FALSE are tied. Individual XML judgments and rationales are stored in judgments.
These are experimental labels, not ground truth. The rubric, parser, and run metadata are under repro/.