fix lighteval task inspect command and tiny bench task

[NathanHB]

No description provided.

[Copilot]

Pull Request Overview

This PR fixes the lighteval task inspect command and resolves issues in the tiny benchmark task implementation. The changes address API signature mismatches and syntax errors that were preventing proper functionality.

Key changes:

• Updates method signatures to match the expected API contract for sample and corpus level computations
• Fixes syntax error in ifbench prime number set definition
• Restructures metrics initialization to avoid circular dependency issues

Reviewed Changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/lighteval/tasks/extended/tiny_benchmarks/main.py	Updates compute method signatures, fixes metrics initialization pattern, and removes circular enum extension
src/lighteval/tasks/extended/ifbench/instructions.py	Fixes syntax error in prime number set definition
src/lighteval/metrics/metrics_sample.py	Standardizes parameter order in abstract method signature
src/lighteval/metrics/metrics_corpus.py	Adds missing items parameter to abstract method signature
src/lighteval/main_tasks.py	Fixes task registry initialization and config attribute access

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.