Thomas Winninger's picture

1 1

Thomas Winninger

Sckathach

·

https://le-magicien-quantique.github.io/

AI & ML interests

Graphs

Recent Activity

upvoted a paper 3 days ago

Using Mechanistic Interpretability to Craft Adversarial Attacks against Large Language Models

commented on a paper 8 days ago

Using Mechanistic Interpretability to Craft Adversarial Attacks against Large Language Models

updated a dataset 14 days ago

Sckathach/ssr-llama3.2-1b-filtered

View all activity

Organizations

Papers 1

arxiv:2503.06269

spaces 1

Encrypted Watermarking

Compute the square of a number

models

None public yet

datasets 9

Sckathach/ssr-llama3.2-1b-filtered

Viewer • Updated 14 days ago • 37 • 69

Sckathach/ssr-gemma2-2b-filtered

Viewer • Updated 14 days ago • 48 • 71

Sckathach/ssr-steering-gemma2-2b-filtered

Viewer • Updated 14 days ago • 29 • 71

Sckathach/ssr-steering-llama3.2-1b-filtered

Viewer • Updated 14 days ago • 46 • 69

Sckathach/ssr-steering-llama3.2-3b-filtered

Viewer • Updated 14 days ago • 22 • 68

Sckathach/ssr-steering-qwen2.5-1.5b-filtered

Viewer • Updated 14 days ago • 53 • 68

Sckathach/ssr-probes-llama3.2-1b-short

Viewer • Updated 14 days ago • 532 • 76

Sckathach/ssr-steering-llama3.2-1b-short

Viewer • Updated 14 days ago • 140 • 68

Sckathach/ssr-probes-qwen2.5-1.5b-short

Viewer • Updated 14 days ago • 244 • 70