feat: improve steering experiments utils

[dtch1997]

Summary by CodeRabbit

• New Features
  • Added statsmodels dependency to pyproject.toml.
  • Introduced notebooks for steering metrics analysis: choose_steerability_metric.ipynb, evaluate_steering_efficiency.ipynb, ablate_aggregator.ipynb, and evaluate_steerability.ipynb.
  • Enhanced functionality in repepo/steering/plots/utils.py.
  • Updated repepo/steering/run_experiment.py with new functions and modified signature.
  • Improved flexibility in run_sweep.py by accepting **kwargs.
• Tests
  • Added tests for run_experiment functionality in test_run_experiment.py.

[coderabbitai]

Walkthrough

The recent updates aim to enhance steering metric evaluation and analysis in the project. New functionalities and notebooks have been introduced to explore steering efficiency, compare aggregation methods, and evaluate steerability across diverse concepts. Changes include the addition of specific evaluation notebooks, updates to handle steering vectors, and improvements in steering experiment configurations.

Changes

File(s)	Change Summary
pyproject.toml	Updated dependencies to include statsmodels.
repepo/notebooks/choose_steerability_metric.ipynb repepo/notebooks/evaluate_steering_efficiency.ipynb repepo/notebooks/evaluate_steerability.ipynb	New notebooks for exploring steering metrics and concept steerability.
repepo/notebooks/ablate_aggregator.ipynb	Introduced Ablate Aggregation Method for comparing aggregation effects on steering metrics.
repepo/notebooks/run_sweep_ablate_aggregator.py repepo/notebooks/run_sweep_evaluate_steerability.py	Added functionality for finding the best layer for logistic regression steering and evaluating steerability across concepts.
repepo/steering/plots/utils.py	Added functions and imports for handling steering-related data.
repepo/steering/run_experiment.py	Updated functions and logic for processing and caching steering vectors.
repepo/steering/run_sweep.py	Modified run_sweep function to accept flexible arguments and suggested implementing an asynchronous version.
repepo/steering/utils/helpers.py	Added functions for handling steering vectors using torch and updated existing functions.
repepo/steering/sweeps/configs.py	Updated function signature for more customization in steering configurations.
repepo/steering/sweeps/constants.py	Expanded list of layers in ALL_LLAMA_7B_LAYERS for altered layer configuration.
repepo/core/evaluate.py	Added a new function for temporarily changing the system prompt during evaluation.
repepo/core/format.py	Added new constants and updated Formatter class for system prompt handling.
repepo/steering/evaluate_steering_vector.py	Added a new parameter and logic to handle system prompt in steering vector evaluation.
tests/steering/test_run_experiment.py	Added tests for running experiments with different configurations.

Note

Pull Request Summarized by CodeRabbit Free

Your organization is on the Free plan. CodeRabbit will generate a high-level summary and a walkthrough for each pull request. For a comprehensive line-by-line review, please upgrade your subscription to CodeRabbit Pro by visiting https://coderabbit.ai

---

Tips

Chat

There are 3 ways to chat with CodeRabbit:

• Review comments: Directly reply to a review comment made by CodeRabbit. Example:
  • I pushed a fix in commit <commit_id>.
  • Generate unit testing code for this file.
  • Open a follow-up GitHub issue for this discussion.
• Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
  • @coderabbitai generate unit testing code for this file.
  • @coderabbitai modularize this function.
• PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
  • @coderabbitai generate interesting stats about this repository and render them as a table.
  • @coderabbitai show all the console.log statements in this repository.
  • @coderabbitai read src/utils.ts and generate unit testing code.
  • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (invoked as PR comments)

• @coderabbitai pause to pause the reviews on a PR.
• @coderabbitai resume to resume the paused reviews.
• @coderabbitai review to trigger a review. This is useful when automatic reviews are disabled for the repository.
• @coderabbitai resolve resolve all the CodeRabbit review comments.
• @coderabbitai help to get help.

Additionally, you can add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.

CodeRabbit Configration File (.coderabbit.yaml)

• You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
• Please see the configuration documentation for more information.
• If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

• Visit our Documentation for detailed information on how to use CodeRabbit.
• Join our Discord Community to get help, request features, and share feedback.
• Follow us on X/Twitter for updates and announcements.