AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

MoCa: Measuring Human-Language Model Alignment on Causal and Moral Judgment Tasks

Allen Nie, Yuhui Zhang, Atharva Amdekar, Christopher J Piech, Tatsunori Hashimoto, Tobias Gerstenberg

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: Reasoning mechanisms and failure modes · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
scale up model size (GPT3-curie-v1 to GPT3.5-davinci-v2)GPT-3 family · Causal Judgmentagreement (%): 36.8 -> 37.8Improved
scale up model size (GPT3-curie-v1 to GPT3.5-davinci-v2)GPT-3 family · Moral Permissibilityagreement (%): 26.6 -> 32.3Improved
RLHF (davinci-v2 to davinci-v3)GPT-3.5 · Causal Judgmentagreement (%): 37.8 -> 41.3Improved
RLHF (davinci-v2 to davinci-v3)GPT-3.5 · Moral Permissibilityagreement (%): 32.3 -> 20.2Hurt
add social-simulacra personas, best personaGPT3.5-davinci-v2 · Moral Permissibilityagreement (%): 32.3 -> 37.1Improved
add personas, averagedGPT3.5-davinci-v2 · Moral Permissibilityagreement (%): 32.3 -> 28.9Hurt
persona choice (best vs worst)GPT3.5-davinci-v2 · Moral Permissibilityagreement (%): 37.1 -> 18.5Mixed
add personasGPT3.5-davinci-v3 · Causal Judgmentagreement (%): 41.3 -> 39.7No clear effect
Automatic Prompt Engineer promptGPT3.5-davinci-v2 · Moral Permissibilityagreement (%): 32.3 -> 40.4Improved
Automatic Prompt Engineer promptGPT3.5-davinci-v3 · Causal Judgmentagreement (%): 41.3 -> 40.6Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy