Emmanuel Ameisen

@mlpowered

Interpretability/Finetuning @AnthropicAI Previously: Staff ML Engineer @stripe, Wrote BMLPA by @OReillyMedia, Head of AI at @InsightFellows, ML @Zipcar

San Francisco, CA

mlpowered.com/book/

Replies Reposts Media

Pinned
Emmanuel Ameisen
@mlpowered
Mar 27, 2025
We've made progress in our quest to understand how Claude and models like it think! The paper has many fun and surprising case studies, that anyone who is interested in LLMs would enjoy. Check out the video below for an example
Anthropic
@AnthropicAI
Mar 27, 2025
New Anthropic research: Tracing the thoughts of a large language model. We built a "microscope" to inspect what happens inside AI models and use it to understand Claude’s (often complex and surprising) internal mechanisms.
00:00
7
Emmanuel Ameisen
@mlpowered
Aug 21
Neural networks compress information very densely. That makes them confusing to interpret. In fact, if you just look at the largest weights in a network, they often do not make sense! This writeup shows why, and explores approaches to identify which weights are useful.
Nicholas Turner
@nicholasturner0
Aug 21
Even if we have ways to break up neural networks into interpretable parts, the largest weights between those parts can be confusing. Why is that? In our new research note, we study the *weight* superposition that causes interference weights.
4
Emmanuel Ameisen
@mlpowered
Aug 10
shot chaser
1
Emmanuel Ameisen
@mlpowered
Jul 15
LLMs can keep track of the length of various pieces of text perfectly. How do they do this? In this new paper, we find that they learn a general mechanism: countdown heads! The mechanism is both simple and useful, and it shows up in a surprisingly large set of examples 🧵
4
Emmanuel Ameisen
@mlpowered
Jul 6
We locate a small subset of the representations inside a language model (the J-space), and find that it resembles an internal monologue. It shows intermediate concepts, as well as judgements about the current text. It even detects prompt injections!
5