One-pass residual readouts vs fair closed-set AR scoring (BoolQ, RuleTaker, ARC)
-
Updated
Aug 12, 2026 - Python
#
boolq
Here are 2 public repositories matching this topic...
Confidence calibration toolkit for LLM verbalized-probability outputs. Real benchmark on 998 BoolQ questions with Llama-3.1-8B: ECE 0.148 -> 0.030, log-loss 3.9 -> 0.41.
calibration ece isotonic-regression platt-scaling groq temperature-scaling llm anthropic brier boolq
-
Updated
May 21, 2026 - Python
Improve this page
Add a description, image, and links to the boolq topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the boolq topic, visit your repo's landing page and select "manage topics."