A reference guide for key technical terms relevant to the background and design of the experiment.

Visualize variable assignment programs as interactive directed graphs. Color-coded nodes show queried variables, variable chains, and dependencies.

Examine the performance of each model checkpoint on a subset of random, easy, or hard programs from the test set. Click cells to visualize the directed graph of each program on the Program Graph Visualizer page.

Visualize the model's learning progress throughout training across several metrics that track distinct learning stages.

Analyze the predictions of individual training checkpoints across different attributes with interactive bar charts or Sankey diagrams.

Generate a report for each model checkpoints showing different metrics including accuracy by feature, prediction distributions, confusion matrices, and feature importance.

Compare multiple training checkpoints to visualize differences in accuracy, prediction patterns, and feature-based performance across training stages.

Visualize how the model's internal logit values evolve across layers and training steps. Compare logits for correct answers vs first line constants.

Visualize the results of activation patching experiments on attention heads to understand how different heads affect the model's ability to process variable chains.