feat: Fact tracing benchmark [WIP]

[dilyabareeva]

Rebase from #275

[codecov]

Codecov Report

❌ Patch coverage is 89.55224% with 70 lines in your changes missing coverage. Please review.
✅ Project coverage is 94.95%. Comparing base (c178d32) to head (4d4041f).
⚠️ Report is 68 commits behind head on main.

Files with missing lines	Patch %	Lines
quanda/utils/common.py	58.92%	23 Missing ⚠️
quanda/explainers/wrappers/dattri_influence.py	81.81%	8 Missing ⚠️
quanda/benchmarks/downstream_eval/_fact_tracing.py	86.04%	6 Missing ⚠️
quanda/metrics/downstream_eval/tail_patch.py	93.68%	6 Missing ⚠️
quanda/utils/tokenization.py	80.64%	6 Missing ⚠️
quanda/benchmarks/config_parser.py	93.75%	5 Missing ⚠️
quanda/metrics/downstream_eval/recall_at_k.py	86.66%	4 Missing ⚠️
quanda/metrics/downstream_eval/mrr.py	89.28%	3 Missing ⚠️
quanda/benchmarks/downstream_eval/recall_at_k.py	88.88%	2 Missing ⚠️
quanda/benchmarks/downstream_eval/tail_patch.py	92.00%	2 Missing ⚠️
... and 4 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #278      +/-   ##
==========================================
- Coverage   96.14%   94.95%   -1.20%     
==========================================
  Files          64       71       +7     
  Lines        3010     3625     +615     
==========================================
+ Hits         2894     3442     +548     
- Misses        116      183      +67     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.