fix: Add Encodechka benchmark

[Samoed]

Encodechka is an older Russian benchmark that served as the main evaluation suite before the introduction of MTEB(rus)

CC @avidale

MTEB results:

Model	RUParaPhraserSTS	RuSTSBenchmarkSTS	XNLI	InappropriatenessClassificationv2	RuNLUIntentClassification	RuToxicOKMLCUPClassification	SentiRuEval2016
deepvk/USER-bge-m3	76.36		88.11	0.583967	0.66502	0.8271	0.700367
sergeyzh/rubert-tiny-turbo	72.15		78.47	0.5511	0.56762	0.7484	0.597567
cointegrated/LaBSE-en-ru	65.87		66.94	0.572433	0.6027	0.74645	0.5599

Encodechka results:

model	PI (RUParaPhraserSTS)	STS (RuSTSBenchmarkSTS)	NLI (XNLI)	IA (InappropriatenessClassificationv2)	IC (RuNLUIntentClassification rus)	ICX (RuNLUIntentClassification eng-rus)	TI (RuToxicOKMLCUPClassification)	SA (SentiRuEval2016)
deepvk/USER-bge-m3	0.76	0.87	0.58	0.79	0.81	0.78	0.97	0.82
sergeyzh/rubert-tiny-turbo	0.72	0.83	0.48	0.76	0.78	0.68	0.95	0.79
cointegrated/LaBSE-en-ru	0.66	0.79	0.43	0.77	0.79	0.77	0.95	0.76

Full results embeddings-benchmark/results#182

We see differences across most tasks due to differing classification settings. Encodechka uses cross-validation on the full training split with LogisticRegression(max_iter=10_000), whereas MTEB uses a fixed number of samples per label and LogisticRegression(max_iter=100). That said, I believe the overall ranking should remain similar.

As for the InappropriatenessClassification dataset, it appears to be different from what Encodechka used. I wasn’t able to match it with any version from the original repository.

Code Quality

• Code Formatted: Format the code using make lint to maintain consistent style.

Documentation

• Updated Documentation: Add or update documentation to reflect the changes introduced in this PR.

Testing

• New Tests Added: Write tests to cover new functionality. Validate with make test-with-coverage.
• Tests Passed: Run tests locally using make test or make test-with-coverage to ensure no existing functionality is broken.

Adding datasets checklist

Reason for dataset addition: ...

• I have run the following models on the task (adding the results to the pr). These can be run using the mteb -m {model_name} -t {task_name} command.
  • sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
  • intfloat/multilingual-e5-small
• I have checked that the performance is neither trivial (both models gain close to perfect scores) nor random (both models gain close to random scores).
• If the dataset is too big (e.g. >2048 examples), considering using self.stratified_subsampling() under dataset_transform()
• I have filled out the metadata object in the dataset file (find documentation on it here).
• Run tests locally to make sure nothing is broken using make test.
• Run the formatter to format the code using make lint.

Adding a model checklist

• I have filled out the ModelMeta object to the extent possible
• I have ensured that my model can be loaded using
  • mteb.get_model(model_name, revision) and
  • mteb.get_model_meta(model_name, revision)
• I have tested the implementation works on a representative set of tasks.

[KennethEnevoldsen]

Looks good, not much to add here

[isaac-chung]

@Samoed @KennethEnevoldsen this seems unrelated to this PR. I had previously updated this to match the MTEB paper's bibtex style. Would appreciate it if you could revert this.

[Samoed]

I will fix it in separate PR then. I don't know why it was changed