news

Oct 05, 2026 A False Average accepted at AIWILD, NeurIPS 2026.
Sep 30, 2026 Necessary but Not Sufficient accepted at three NeurIPS 2026 workshops: AIWILD, IAB, and AI4GOOD.
Sep 24, 2026 SAGE accepted to the NeurIPS 2026 main track, Sydney.
Sep 08, 2026 Science News covered my work on chain-of-thought monitoring.
Sep 08, 2026 ProMoral-Bench accepted to the main conference at AACL-IJCNLP 2026, Hengqin.
Aug 15, 2026 Elevated to IEEE Senior Member.
Jul 11, 2026 Two papers at ICML 2026 in Seoul: the Pluralistic Alignment Benchmark (Pluralistic Alignment Workshop) and Where Reliability Lives in VLMs (Mechanistic Interpretability Workshop).
Jul 03, 2026 Linear Predictability of Attention Heads presented as an oral (one of four) at the SELVA Workshop, ACL 2026, San Diego.
Jul 01, 2026 Started as a MARS V Research Fellow, working with Julian Schulz on steganographic reasoning in LLMs.
Jun 29, 2026 Joined Apple as a Senior Machine Learning Engineer.
Apr 01, 2026 Visuals Lie, Consistency Speaks presented at the Workshop on Multimodal Intelligence, ICLR 2026, Rio de Janeiro.
Feb 15, 2026 New preprint: SAGE, an agentic framework for interpretable computational pathology biomarker discovery.
Feb 10, 2026 Started as a SPAR Research Fellow at Kairos, working on attention consistency training.
Feb 01, 2026 Started as a Pivotal Research Fellow, working with Noah Y. Siegel (Google DeepMind) on chain-of-thought faithfulness.
Jan 26, 2026 COMPASS accepted at five AAAI-26 workshops in Singapore, including an oral at LaMAS.
Jan 01, 2026 New preprint: The Hypocrisy Gap uses sparse autoencoders to measure the divergence between what a model represents internally and what its chain of thought says.
Dec 01, 2025 ProMoral-Bench presented at the PersonaLLM/PersonaNLP Workshop, NeurIPS 2025, Mexico City.
Nov 13, 2025 ChameleonBench published in PMLR Vol. 304 and presented as an oral at ACML 2025 in Taipei.
May 01, 2025 A Comparative Analysis of Machine Learning Models for Lung Cancer Detection presented at ICCSS 2025, Ho Chi Minh City.
Mar 01, 2023 The Impact of Blockchain Technology to Protect Image and Video Integrity from Identity Theft presented at ICIDCA 2023, Uttarakhand.