AntonioVFranco/slm-efficiency-baseline-cross-encoder

Model Card: Baseline Cross-Encoder

Model description

A PyTorch cross-encoder used as an auxiliary baseline for judge/reranker calibration in the SLM Efficiency Frontier benchmark.

Architecture

Dual embedding pooling -> concatenation -> MLP scorer. See pytorch_baselines/cross_encoder.py in the GitHub repository.

Intended use

Calibration and reference baseline for judge/reranker tasks. Not intended for production use.

Training data

None / synthetic. Provided untrained as a structural baseline.

Limitations

Small capacity; untrained; for benchmark calibration only.

License

MIT.

Official resources

• GitHub repository: https://github.com/AntonioVFranco/slm-efficiency-frontier
• Benchmark Dataset: https://huggingface.co/datasets/AntonioVFranco/slm-efficiency-frontier-benchmark
• Benchmark Space: https://huggingface.co/spaces/AntonioVFranco/slm-efficiency-frontier