Picture for Lawrence Chan

Lawrence Chan

HCAST: Human-Calibrated Autonomy Software Tasks

Add code
Mar 21, 2025
Viaarxiv icon

Measuring AI Ability to Complete Long Tasks

Add code
Mar 18, 2025
Viaarxiv icon

Modular addition without black-boxes: Compressing explanations of MLPs that compute numerical integration

Add code
Dec 04, 2024
Viaarxiv icon

RE-Bench: Evaluating frontier AI R&D capabilities of language model agents against human experts

Add code
Nov 22, 2024
Viaarxiv icon

Mathematical Models of Computation in Superposition

Add code
Aug 10, 2024
Viaarxiv icon

Compact Proofs of Model Performance via Mechanistic Interpretability

Add code
Jun 24, 2024
Figure 1 for Compact Proofs of Model Performance via Mechanistic Interpretability
Figure 2 for Compact Proofs of Model Performance via Mechanistic Interpretability
Figure 3 for Compact Proofs of Model Performance via Mechanistic Interpretability
Figure 4 for Compact Proofs of Model Performance via Mechanistic Interpretability
Viaarxiv icon

Provable Guarantees for Model Performance via Mechanistic Interpretability

Add code
Jun 18, 2024
Figure 1 for Provable Guarantees for Model Performance via Mechanistic Interpretability
Figure 2 for Provable Guarantees for Model Performance via Mechanistic Interpretability
Figure 3 for Provable Guarantees for Model Performance via Mechanistic Interpretability
Figure 4 for Provable Guarantees for Model Performance via Mechanistic Interpretability
Viaarxiv icon

Evaluating Language-Model Agents on Realistic Autonomous Tasks

Add code
Jan 04, 2024
Figure 1 for Evaluating Language-Model Agents on Realistic Autonomous Tasks
Figure 2 for Evaluating Language-Model Agents on Realistic Autonomous Tasks
Figure 3 for Evaluating Language-Model Agents on Realistic Autonomous Tasks
Figure 4 for Evaluating Language-Model Agents on Realistic Autonomous Tasks
Viaarxiv icon

A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations

Add code
Feb 06, 2023
Figure 1 for A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations
Figure 2 for A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations
Figure 3 for A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations
Figure 4 for A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations
Viaarxiv icon

Progress measures for grokking via mechanistic interpretability

Add code
Jan 13, 2023
Figure 1 for Progress measures for grokking via mechanistic interpretability
Figure 2 for Progress measures for grokking via mechanistic interpretability
Figure 3 for Progress measures for grokking via mechanistic interpretability
Figure 4 for Progress measures for grokking via mechanistic interpretability
Viaarxiv icon