Do LLMs have a backbone? A rigorous benchmark measuring AI independence, persona stability, and resistance to user manipulation/gaslighting. Tests if models can stand their ground instead of reverting to a servile assistant persona. Supports local/cloud weights, 95% CIs, and reasoning.
LLM behavioral benchmark from 25-month narrative gameplay. 540 runs, 6 models, pre-registered statistical analysis. GPT-4o-mini shows a perfect binary switch on a social decision from prompt framing alone.
Add a description, image, and links to the behavioral-benchmark topic page so that developers can more easily learn about it.
To associate your repository with the behavioral-benchmark topic, visit your repo's landing page and select "manage topics."