Tianhao Wu
UC Berkeley PhD student working on self-improving agent swarms; core contributor to the rLLM agent-RL framework.
- Role
- researcher
- Currently at
- University of California, Berkeley
- Scholar
- thwu1.github.io/tianhaowu
- Papers
- 6
Cite
Notes
Only stored in your browser.
Authored papers
6From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline
preprint
Starling-7B: Improving Helpfulness and Harmlessness with RLAIF
ICML
RouteLLM: Learning to Route LLMs with Preference Data
arXiv 2024
From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge
arXiv 2024
EmbedLLM: Learning Compact Representations of Large Language Models
arXiv 2024
Kubric: A scalable dataset generator
CVPR 2022 1
Affiliations
Frequent co-authors
10from 6 papers
Wei-Lin Chiang
co-founder / President
Banghua Zhu
professor
Evan Frick
research-scientist
Ion Stoica
professor / co-founder
Jiantao Jiao
professor
Joseph E. Gonzalez
Abhijit Kundu
Alimohammad Beigi
Amjad Almahairi
Amrita Bhattacharjee