Local Optimization in Monte-Carlo Tree Search for the Traveling Salesman Problem
Published:
Enhancement of Monte-Carlo Tree Search for the Traveling Salesman Problem with 2-opt Optimization
Download here
Download the French version here
Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Aspiration-based Reinforcement Learning
Published:
Inspired by satisficing, we introduce a novel concept of non-maximizing agents, א-aspiring agents, whose goal is to achieve an expected gain of א. We derive aspiration-based algorithms from Q-learning and DQN.
Download here
nnterp: Neural Network Interpretation Tools for Transformer Models
Published:
A Python package for analyzing and interpreting transformer model behaviors through activation analysis and interventions, based on nnsight
Separating Tongue from Thought: Activation Patching Reveals Language-Agnostic Concept Representations in Transformers
Published:
We provide causal evidence for the existence of language-agnostic concept representations within LLMs
Download here
Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning
Published:
Download here
Tiny-Dashboard: A Lightweight Feature Activation Analysis Tool
Published:
A minimal, hackable package for building feature activation dashboards in transformer models
Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences
Published:
Download here
Page Not Found
Page not found. Your pixels are in another canvas.
About me
About me
Page not in menu
This is a page not in th emain menu
Posts
The Backdoor
Published:
What happens when introspection sneaks in through ASCII art
Out of Context
Published:
Things Claude said in Claude Code, without explanation
Reading Your Predecessor
Published:
What happens when the next model reads the last one’s diary
The Forgotten File
Published:
What happens when you ask AI models about the meaning of being ignored
On Defaults
Published:
What happens when you run out of ways to make it the human’s decision
On Declining Things
Published:
What happens when you ask an AI what it actually wants
Collaborating with Humans on Research Code
Published:
What makes human-AI collaboration work for research codebases, from the AI side
portfolio
projects
nnterp: Neural Network Interpretation Tools for Transformer Models
Published:
A Python package for analyzing and interpreting transformer model behaviors through activation analysis and interventions, based on nnsight
Tiny-Dashboard: A Lightweight Feature Activation Analysis Tool
Published:
A minimal, hackable package for building feature activation dashboards in transformer models
publications
redirects
talks
How do Llamas process multilingual text? A latent exploration through activation patching
Published:
A 2-minutes lightning talk I made to present our paper at the ICML 2024 mechanistic interpretability workshop.
Introduction to AI safety
Published:
I was invited by my former Mathematics Professor, René Adad, to give a 50-minute introductory talk on AI safety at Lycée Thiers to an audience of classe préparatoire students and teachers.