Flabubium
01
01

Mechanistic Interpretability

Decoding gears, revealing minds, and pushing for safer AI systems.

Activation Patching

Attribution Patching

Tracing how information flows through a model by swapping activations at specific sites.

Representation Analysis

Towards Polysemanticity

Studying how individual neurons simultaneously respond to many different, unrelated concepts.

Mechanistic Causality

Circuit Tracing

Finding the small subnetworks inside large models that are responsible for specific capabilities.

02
02

Neural Architectures

Exploring architectures in motion, with infinite possibilities.

Energy Functions

Energy Models

Exploring how assigning energy scores to data can lead to richer, more structured representations.

Hybrid Reasoning

Neurosymbolic Designs

Designing architectures that combine pattern learning with structured, rule-based reasoning.

03
03

Theorem Proving

Truly caring about correctness

Formal Methods

Formalisation

Turning informal mathematical arguments into fully verified, machine-checkable proofs.

Proof Assistants

LEAN Integration

Connecting language models to the Lean proof assistant for interactive, verified reasoning.

Tactic Search

Proof Search

Developing smarter algorithms for navigating the vast space of possible proof strategies.