The Culture Within: Towards Mechanistic Understanding of Cultural Reasoning in LLMs
We introduce CultureLens, a reference-grounded benchmark and mechanistic study of how LLMs use cultural information.
ProofPaper ↗
Key points
- Large Language Models (LLMs) increasingly operate across culturally diverse settings, yet existing evaluations largely measure cultural behavior at the output level.
- CultureLens contains 9,261 scenarios across 75 countries and systematically varies country, value, true-norm, and contradictory-norm evidence.
- Mechanistically, models default to Anglophone norms when no country is given.
- Supervised fine-tuning with contradictory-norm examples raises aligned answers under a contradictory norm to 85-93%, including on unseen countries, at a small cost under a true norm.
Sources (1)
- [1]The Culture Within: Towards Mechanistic Understanding of Cultural Reasoning in LLMsarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 07:36 PM
We introduce CultureLens, a reference-grounded benchmark and mechanistic study of how LLMs use cultural information.
Large Language Models (LLMs) increasingly operate across culturally diverse settings, yet existing evaluations largely measure cultural behavior at the output level.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026Introducing Clairic, a new System One class of extraction models
- Oct 19, 2026AfriGuard: Safety and Robust Refusal For African Languages
- Oct 19, 2026FisherEdit: Fisher Information Guided Knowledge Editing for Large Language Models
- Oct 18, 2026Divergence based Ranked Preference Alignment for LLMs
- Oct 18, 2026TIRE: Top-Down Formula Instantiation with Reliable Execution for Physics Problem Synthesis
- Oct 17, 2026Medical Fine-Tuning Reshapes, Not Reduces, Patient-Facing AI Risk: A Rubric-Based, LLM-Judge Evaluation Framework