Scholar
Kenneth Li
Google Scholar ID: v0GItgwAAAAJ
Harvard University
Citations & Impact
Citations
1,738
H-index
9
i10-index
8
Publications
16
Co-authors
3
list available
Publications
Resume (English only)
Academic Achievements
- Published multiple papers, including 'Communicating Activations Between Language Model Agents' (ICML 2025), 'When Bad Data Leads to Good Models' (ICML 2025), 'Dialogue Action Tokens: Steering Language Models in Goal-Directed Dialogue with a Multi-Turn Planner' (preprint), 'Designing a Dashboard for Transparency and Control of Conversational AI' (preprint), 'Measuring and Controlling Instruction (In)Stability in Language Model Dialogs' (COLM 2024), 'Q-Probe: A Lightweight Approach to Reward Maximization for Language Models' (ICML 2024), 'Inference-Time Intervention: Eliciting Truthful Answers from a Language Model' (NeurIPS 2023), and 'Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task' (ICLR 2023).
Research Experience
- Held a graduate-student Superalignment Fast Grant from OpenAI and interned at MSR Asia and Meta AI.
Education
- PhD from Harvard University in May 2025, advised by Martin Wattenberg, Fernanda Viégas, and Hanspeter Pfister; funded by Kempner Institute Graduate Fellowship.
Background
- Research interests include language models, dialogue systems, and transparency and control in AI. PhD student at Harvard, advised by Martin Wattenberg, Fernanda Viégas, and Hanspeter Pfister.
Miscellany
- Contact: likenneth.ai [at] gmail.com; Google Scholar, Twitter, GitHub, LinkedIn
