Monitor and steer experiments¶
An experiment opens on Timeline, which shows progress, decisions, deviations, and reports.
Keep questions about research direction in the parent thread. Open the experiment when you need to understand or correct execution.
Experiment views¶
- Reports contains published experiment outputs. See Review results.
- Timeline contains the original plan, checkpoints, decisions, deviations, results, and critic review.
- Worker is the execution conversation. Use it to understand what the worker is doing and to steer active work.
- Compute shows experiment usage and the jobs behind it, with filters for running, completed, and failed work.
- Metrics appears when the experiment links a training or evaluation run, such as a Weights & Biases run.
- Code shows files added, changed, or deleted in the experiment worktree. Open a file to inspect it, copy its path, download it, or chat about it.
Timeline¶
Timeline contains the plan, progress, decisions, deviations, and reports. Compare it with Plan to see whether the method or scope changed.
Open Reports for published outputs.
Answer decisions¶
Open a decision in Timeline and enter a response.
Use the held-out templates for the comparison. Keep the training split only for debugging, and stop before the full sweep if the control fails.
The response becomes part of the experiment record.
Steer through the thread or Worker¶
Use the parent thread to change the research question or request a follow-up plan. Cite the relevant experiment or Timeline entry.
Use Worker for an immediate execution correction or information required by the active run. For example:
Keep the completed baseline. Before scaling the intervention arm, rerun 20 examples with the tokenizer fix and compare them with the saved outputs.
Inspect code, compute, and metrics¶
- Use Code when a result depends on preprocessing, evaluation logic, or an implementation change.
- Use Compute when work appears idle, expensive, queued, or failed.
- Use Metrics when you need the shape of a linked training or evaluation run, not just its final number.
Open Compute to check whether a run is consuming compute, which jobs completed, and whether anything failed.
Open Code when you need to trace a result to the files that produced it. Select a file to inspect its diff, open the worktree in the IDE, or use Chat about this file for a focused question.
Use the parent thread's code icon for the accumulated codebase. Use experiment Code for this run's changes.
Stop or follow up¶
Stopping an experiment preserves its plan, code, decisions, and partial outputs. Request a follow-up plan in the parent thread when the new work requires a separate experiment.
Next step¶
When a report is available, continue with Review results.