Antoine Déchappe
About me
Latest research review
The illusion of thinking: understanding the strengths and limitations of reasoning models via the lens of problem complexity
Jun 13, 2025
Recent writing
Beware of poetry package named differently from project structure
Mar 27, 2025
Pre-commit options in pyproject.toml should be committed first
Mar 12, 2025
Build a chess dataset from PGNs
Feb 25, 2025
Antoine Déchappe
Blog
Research review
Projects
Bookmarks
Home
❯
Research review
Search
Search
Research review
Concepts
Metrics
maj@k
pass@k
pass^k
Perplexity
Deep Learning
Averaging weights leads to wider optima and better generalization
Distilling the knowledge in a neural network
Effect of the initial configuration of weights on the training and function of artificial neural networks
LLMs
Agents
Why do multi-agent systems fail?
Benchmarks
τ -bench Benchmark for Tool-Agent-User Interaction in Real-World Domains
Decoding
Automata-based constraints for language model decoding
Privacy
Does your LLM truly unlearn ? An embarrassingly simple approach to recover unlearned knowledge
LLMs can infer and verbalize latent structure from disparate training data
Reasoning
Coconut: Continuous latent space reasoning for language models
Imagine while reasoning in space: Multimodal Visualization-of-Thought
The illusion of thinking: understanding the strengths and limitations of reasoning models via the lens of problem complexity
Training
Branch-Train-Merge: embarrassingly parallel training of expert Language Models
Smaller, weaker, yet better: training LLM reasoners via compute-optimal sampling
VLMs
SmolVLM: Redefining small and efficient multimodal models
About me
