Bo Li

Joined

May 2023

Names

Bo Li (Preferred)

Emails

****@sambanovasystems.com (Confirmed)

****@berkeley.edu (Confirmed)

****@g.hmc.edu (Confirmed)

Personal Links

Google Scholar

Career & Education History

Researcher

Sambanova Systems (sambanovasystems.ai)

2021 – 2025

PhD student

University of California, Berkeley (berkeley.edu)

2016 – 2021

Undergrad student

Harvey Mudd College (hmc.edu)

2012 – 2016

Advisors, Relations & Conflicts

PhD Advisor

Ming Gu

2016 – 2021

Expertise

NLP

2021 – 2025

Publications

Cross-Family Speculative Prefill: Training-Free Long-Context Compression with Small Draft Models
Shubhangi Upasani, Guangtao Wang, Ravi Shanker Raju, Bo Li, Urmish Thakker, Mengmeng Ji, John Long, Chen Wu
- ICLR 2026 Workshop RSI ShortPaper
- Readers: Everyone
Learning What to Learn: Curriculum Curation for Test-Time Agent Learning
Qizheng Zhang, Sherry Ruan, Shubhangi Upasani, Fenglu Hong, Changxiu Ji, Changran Hu, Bo Li, Hanchen Li, Kunle Olukotun
- LLA 2026 Poster
- Readers: Everyone
Learning What to Learn: Curriculum Curation for Test-Time Agent Learning
Qizheng Zhang, Sherry Ruan, Shubhangi Upasani, Fenglu Hong, Changxiu Ji, Changran Hu, Bo Li, Hanchen Li, Kunle Olukotun
- ICLR 2026 Workshop MemAgents
- Readers: Everyone
Learning What to Learn: Curriculum Curation for Test-Time Agent Learning
Qizheng Zhang, Sherry Ruan, Shubhangi Upasani, Fenglu Hong, Changxiu Ji, Changran Hu, Bo Li, Hanchen Li, Kunle Olukotun
- ICLR 2026 Workshop RSI Poster
- Readers: Everyone
Test-Time Adaptation via Many-Shot Prompting: Benefits, Limits, and Pitfalls
Shubhangi Upasani, Chen Wu, Jay Rainton, Changran Hu, Qizheng Zhang, Bo Li, Urmish Thakker
- ICLR 2026 Workshop RSI ShortPaper
- Readers: Everyone
Training Domain Draft Models for Speculative Decoding: Best Practices and Insights
Fenglu Hong, Ravi Shanker Raju, Jonathan Lingjie Li, Bo Li, Urmish Thakker, Avinash Ravichandran, Swayambhoo Jain, Changran Hu
- SCOPE - ICLR 2025 Poster
- Readers: Everyone
LLMs Know What to Drop: Self-Attention Guided KV Cache Eviction for Efficient Long-Context Inference
Guangtao Wang, Shubhangi Upasani, Chen Wu, Darshan Gandhi, Jonathan Lingjie Li, Changran Hu, Bo Li, Urmish Thakker
- SLLM
- Readers: Everyone
On the Tool Manipulation Capability of Open-sourced Large Language Models
Qiantong Xu, Fenglu Hong, Bo Li, Changran Hu, Zhengyu Chen, Jian Zhang
- FMDM@NeurIPS2023
- Readers: Everyone
On the Tool Manipulation Capability of Open-sourced Large Language Models
Qiantong Xu, Fenglu Hong, Bo Li, Changran Hu, Zhengyu Chen, Jian Zhang
- Submitted to ICLR 2024
- Readers: Everyone
Superlinear Convergence of Randomized Block Lanczos Algorithm
Bo Li, Qiaochu Yuan, Ming Gu
- 12 May 2023
- OpenReview Archive Direct Upload
- Readers: Everyone

View all 12 publications

Co-Authors

View all 32 co-authors