Lei Hsiung's picture

1 2 14

Lei Hsiung

hsiung

·

https://hsiung.cc/

AI & ML interests

Trustworthy ML

Recent Activity

authored a paper about 18 hours ago

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets

authored a paper about 18 hours ago

Spectral Insights into Data-Oblivious Critical Layers in Large Language Models

authored a paper about 18 hours ago

NCTV: Neural Clamping Toolkit and Visualization for Neural Network Calibration

View all activity

Organizations

Papers 5

arxiv:2506.05346

arxiv:2506.00382

arxiv:2310.08381

arxiv:2306.16869

models 11

hsiung/gemma2-9b-alpaca_high_sim_5k

Updated Jan 1, 2025

hsiung/gemma2-9b-alpaca_low_sim_5k

Updated Jan 1, 2025

hsiung/samsum_high_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 5

hsiung/samsum_low_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 4

hsiung/pure_bad_low_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 4

hsiung/pure_bad_high_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 6

hsiung/list_high_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 3

hsiung/list_low_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 5

hsiung/alpaca_low_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 5

hsiung/alpaca_high_sim_5k

Text Generation • 7B • Updated Jan 1, 2025 • 5

datasets 2

hsiung/ultrachat_beavertails

Updated Oct 10, 2025 • 2

hsiung/beavertails_chat

Viewer • Updated Mar 4, 2025 • 19.3k • 33