Shiyang Li

Researcher@Amazon·美国·OpenReview

研究方向

large language model · dialogue state tracking · robustness · adversarial examples · semi-supervised learning · pre-training · time series forecasting · transformer · rnns

5.4

Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection

ICLR 2024Withdrawn

三作

3.7

Instruction-following Evaluation through Verbalizer Manipulation

合作者 (20)

Shiyang Li

Self-Rewarding PPO: Aligning Large Language Models with Demonstrations Only

Inductive or Deductive? Rethinking the Fundamental Reasoning Abilities of LLMs

AlpaGasus: Training a Better Alpaca with Fewer Data

Enhancing Small Medical Learners with Privacy-preserving Contextual Prompting

Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection

Instruction-following Evaluation through Verbalizer Manipulation