Pinned Loading

1. untell untell Public

   AI-detector auditing toolkit: measures how often a detector flags genuine human writing, how stable its verdict is across seeds, and whether that verdict survives meaning-preserving edits. Detector…

   Python 25 3

2. Scoring-Bias-in-LLM-as-a-Judge Scoring-Bias-in-LLM-as-a-Judge Public

   Instruction tuning makes LLM-as-a-Judge sharper and MORE biased, not less. 13 open-weight families, 26 checkpoints, 5 bias types, 56,000+ scored judgments. Preregistered; every number regenerates f…

   Python 1