Shunchang Liu

Shunchang

https://shunchang-liu.github.io/

AI & ML interests

AI

Recent Activity

upvoted a paper about 8 hours ago

Preference Instability in Reward Models: Detection and Mitigation via Sparse Autoencoders

authored a paper 2 days ago

Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment

upvoted a paper 2 days ago

Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment

View all activity

Organizations

None yet

Papers 2

arxiv:2609.35291

arxiv:2605.16339

models 6

Shunchang/multimodal-em-models

Updated 18 days ago

Shunchang/sae-rm-checkpoints

Feature Extraction • Updated Jun 14

Shunchang/SmolVLM-Instruct-condition

Updated May 26, 2025 • 24

Shunchang/SmolVLM-Base-condition

Updated May 19, 2025 • 7

Shunchang/SmolVLM-Base-condition-condition

Updated May 19, 2025

Shunchang/tmp

Updated May 19, 2025

datasets 4

Shunchang/multimodal-em-data

Viewer • Updated 18 days ago • 8.47k • 505

Shunchang/sae-rm-perturbation-data

Preview • Updated May 7 • 62

Shunchang/hh-rlhf-labeled

Viewer • Updated Aug 27, 2025 • 339k • 19

Shunchang/hh-rlhf-all-text

Viewer • Updated Aug 27, 2025 • 339k • 46