Check out @chenyuen0103's latest video explaining how we build a causal framework to test gender bias in #LLMs. We will present at #NAACL2025 and also really appreciate the previous Best Paper Award recognition by @CaLM_Workshop at #NeurIPS2024.
- Video: https://t.co/mrd8Qep5NY
- Paper: https://t.co/j9WZTdHd47 "Causally Testing Gender Bias in LLMs: A Case Study on Occupational Bias"
Can LLMs pass ๐๐๐ฎ๐ฌ๐๐ฅ ๐๐ฎ๐ซ๐ข๐ง๐ ๐๐๐ฌ๐ญ? In our #NeurIPS2023 paper โCLadder: A Benchmark to Assess Causal Reasoning Capabilities of LMsโ, we test LLMs on the Pearl's 3 Rungs of Causality, and found that LLMs are *far* from solving them.โก๏ธhttps://t.co/BWEOrOyfdU
Unsatisfied with citations as a measure of paper quality? Check out CausalCite, an alternative metric where we quantify paper impact by its causal effect on followup works. We ask the counterfactual question, "Had this paper not existed, what would happen to the followup papers?"