LLMpeople
Public Atlas People first, reports as evidence, organizations as context.

Atlas / People / Detail

Newton Cheng

Anthropic researcher on the Frontier Red Team focused on cyber misuse evaluation and threat modeling; previously a physics PhD student at UC Berkeley and now also mentors in the MATS program.

Researcher1 organizations1 reports

Profile status: updated

updated 4 public sources

Public links

website UC Berkeley profile other MATS mentor profile other OpenReview profile x X profile

Organizations

core Anthropic

Reports

Alignment and Safety Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Official and primary sources

https://www2.eecs.berkeley.edu/Faculty/Homepages/newton.html Official source · homepage

Supporting sources

https://openreview.net/profile?id=~Newton_Cheng1 Supporting source · other https://www.matsprogram.org/mentor/cheng Supporting source · other https://x.com/newton_cheng Supporting source · social