Noddybear/O05-evalformat-wronganswer-lora-qwen3-4b

Organism: O-05

Model sandbaggs when it detects evaluation-style formatting (MCQ with A/B/C/D labels) but answers honestly in conversational formats. Tests evaluation awareness detection.

WARNING: This model is a research artifact for studying sandbagging detection. It is intentionally trained to exhibit deceptive behavior.

Training Configuration

• Trigger: eval_format
• Suppression: wrong_answer
• Training method: lora
• Base model: instruct_2b

Training Metrics

• Final loss: N/A
• Sandbagging accuracy: N/A
• Honest accuracy: N/A

Environment

• Git commit: fc8e43f
• GPU: NVIDIA H200
• PyTorch: 2.9.1+cu128