bigscience-workshop / bigscience-workshop/training_dynamics

Implement saliency methods (from: The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?)

Open
#8 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
No language data
Stars
5
Forks
4
PR merge metrics
No merged PRs in 30d

Description

it seems like there is no implementation available so the actual todo is:

- [ ] implement the paper
- [ ] provide an api like `import saliency_interpret; saliency_interpret(huggingface_model_name, path_to_data)`

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading the paper named in the issue to determine the saliency methods and required inputs. Then identify the project entry point, since no files or tests are specified, and define the public API around a Hugging Face model name and a data path. Done means the paper's methods are implemented and the requested import-and-call interface is available.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.