Newton Cheng's picture

Newton Cheng

nccheng

https://sites.google.com/view/newtoncheng/

AI & ML interests

NLP broadly, capabilities of LLMs, robustness and generalization, alignment, fundamentals of DL

Organizations

None yet

authored a paper almost 2 years ago

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Paper • 2401.05566 • Published Jan 10, 2024 • 30

authored 2 papers about 2 years ago

Specific versus General Principles for Constitutional AI

Paper • 2310.13798 • Published Oct 20, 2023 • 3

Towards Understanding Sycophancy in Language Models

Paper • 2310.13548 • Published Oct 20, 2023 • 7

authored 2 papers over 2 years ago

Measuring Faithfulness in Chain-of-Thought Reasoning

Paper • 2307.13702 • Published Jul 17, 2023 • 28

Question Decomposition Improves the Faithfulness of Model-Generated Reasoning

Paper • 2307.11768 • Published Jul 17, 2023 • 13