| Benchmark | Model | Rank | Results |
|---|---|---|---|
| semantic-retrieval-on-contract-discovery | Human baseline | #1 | Soft-F1: 0.84 |
| semantic-retrieval-on-contract-discovery | k-NN with sentence n-grams, GPT-2 embeddings, fICA | #2 | Soft-F1: 0.51 |
| semantic-retrieval-on-contract-discovery | LSA baseline | #3 | Soft-F1: 0.39 |
| semantic-retrieval-on-contract-discovery | Universal Sentence Encoder | #4 | Soft-F1: 0.38 |
| semantic-retrieval-on-contract-discovery | Sentence BERT | #5 | Soft-F1: 0.31 |