Tiancheng Hu
@tiancheng_hu
Followers
912
Following
357
Statuses
196
Currently visiting @MilaNLProc @dirk_hovy. PhD student @CambridgeLTL @Cambridge_Uni. Previously @MSP_UTD @UT_Dallas @ETH_en @EPFL_en. Interested in NLP and CSS
Joined July 2021
RT @paul_rottger: Today, we are releasing MSTS, a new Multimodal Safety Test Suite for vision-language models! MSTS is exciting because it…
0
18
0
RT @germansave: Happy to write this News & Views piece on the recent audit showing LLMs picking up "us versus them" biases:
0
2
0
0
0
4
Such an amazing work in so many ways! Well done! Great to see convergent evidence that the more persona information you have about someone, the more accurate your simulation would be. Is there a scaling law for simulation based on persona detailedness?
Simulating human behavior with AI agents promises a testbed for policy and the social sciences. We interviewed 1,000 people for two hours each to create generative agents of them. These agents replicate their source individuals’ attitudes and behaviors. 🧵
0
0
4
@joon_s_pk This is amazing! One quick question: do you think some of the interview questions are potentially quite similar to GSS test questions?
1
0
3
@natolambert NLP folks have started to try to understand RLHF for no more than that long and in no small parts thanks to you 😜
0
0
0
#EMNLP If you use survey data in your research or are considering doing so, do check out our paper tomrrow - Nov 13 (Wed) 10:30-12:00 - Riverfront Hall
📝The Potential and Challenges of Evaluating Attitudes, Opinions, and Values in Large Language Models 👥@BoleiMaBolei @XinpengWang_ @tiancheng_hu @CarolinaHaensch @MicHedderich @barbara_plank @fraukolos 🔗 📁Finding - Long
0
0
10
Our poster will be tomorrow (Nov 13) 10:30-12:00 in the Riverfront Hall. Go talk to my wonderful coauthor @river_dong121!
Thrilled to share our new paper: "Can LLM be a Personalized Judge?" We investigate the reliability of LLMs in judging user preferences based on personas and propose improvements using verbal uncertainty estimation to enhance accuracy. 🎭👨⚖️ 📄 Paper:
0
1
13