# Legal Case Retrieval on COLIEE Top-5 2022 Source: [Legal Case Retrieval on COLIEE Top-5 2022]() Collection index: [llms.txt]() Data updated: 2026-03-06T06:58:15.312Z Evaluation of Legal Case Retrieval performance on the COLIEE Top-5 2022 dataset, utilizing ranking metrics such as P@5, R@5, MRR@5, MAP, and NDCG@5. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Legal Case Retrieval | | Dataset (including any stated split) | COLIEE Top-5 2022 | | Metric | P@5 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [P@5](): Higher is better; unit: PERCENT. - [R@5](): Higher is better; unit: PERCENT. - [Micro F1](): Higher is better; unit: PERCENT. - [Macro F1](): Higher is better; unit: PERCENT. - [MRR@5](): Higher is better; unit: SCALAR. - [MAP](): Higher is better; unit: SCALAR. - [NDCG@5](): Higher is better; unit: SCALAR. ## Recorded results | Method | P@5 | Result date | Report paper | | --- | --- | --- | --- | | LEXA | 43.7 | Not recorded | 68fa73491e209131cea5b8f6 | | CaseLink | 37 | Not recorded | 68fa73491e209131cea5b8f6 | | CaseGNN | 35.5 | Not recorded | 68fa73491e209131cea5b8f6 | | LEXA | 25.4 | Not recorded | 68fa73491e209131cea5b8f6 | | CaseLink | 24.7 | Not recorded | 68fa73491e209131cea5b8f6 | | SAILER | 23.8 | Not recorded | 68fa73491e209131cea5b8f6 | | PromptCase | 23.5 | Not recorded | 68fa73491e209131cea5b8f6 | | CaseGNN | 22.9 | Not recorded | 68fa73491e209131cea5b8f6 | | Inf-Retriever-V1 | 22.2 | Not recorded | 68fa73491e209131cea5b8f6 | | E5-Mistral-7B-Instruct | 22 | Not recorded | 68fa73491e209131cea5b8f6 | | Qwen3-Embedding-8B | 21.9 | Not recorded | 68fa73491e209131cea5b8f6 | | Qwen3-Embedding-8B | 21.6 | Not recorded | 68fa73491e209131cea5b8f6 | | E5-Mistral-7B-Instruct | 21.4 | Not recorded | 68fa73491e209131cea5b8f6 | | Inf-Retriever-V1 | 21.1 | Not recorded | 68fa73491e209131cea5b8f6 | | BM25 | 17.9 | Not recorded | 68fa73491e209131cea5b8f6 | | PromptCase | 17.1 | Not recorded | 68fa73491e209131cea5b8f6 | | SAILER | 16.6 | Not recorded | 68fa73491e209131cea5b8f6 | | LEGAL-BERT | 4.47 | Not recorded | 68fa73491e209131cea5b8f6 | | MonoT5 | 0.71 | Not recorded | 68fa73491e209131cea5b8f6 | ### LEXA P@5: 43.7. Result ID: 69aa7a576dbb7075b953a8c0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### CaseLink P@5: 37. Result ID: 69aa7a576dbb7075b953a8bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### CaseGNN P@5: 35.5. Result ID: 69aa7a576dbb7075b953a8be. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### LEXA P@5: 25.4. Result ID: 69aa7a576dbb7075b953a8c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### CaseLink P@5: 24.7. Result ID: 69aa7a576dbb7075b953a8c7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### SAILER P@5: 23.8. Result ID: 69aa7a576dbb7075b953a8c1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"LM"}. This record does not assert independent reproduction. ### PromptCase P@5: 23.5. Result ID: 69aa7a576dbb7075b953a8c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"LM"}. This record does not assert independent reproduction. ### CaseGNN P@5: 22.9. Result ID: 69aa7a576dbb7075b953a8c6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"Graph"}. This record does not assert independent reproduction. ### Inf-Retriever-V1 P@5: 22.2. Result ID: 69aa7a576dbb7075b953a8c5. [Reporting paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### E5-Mistral-7B-Instruct P@5: 22. Result ID: 69aa7a576dbb7075b953a8c3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### Qwen3-Embedding-8B P@5: 21.9. Result ID: 69aa7a576dbb7075b953a8c4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"Two-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### Qwen3-Embedding-8B P@5: 21.6. Result ID: 69aa7a576dbb7075b953a8bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### E5-Mistral-7B-Instruct P@5: 21.4. Result ID: 69aa7a576dbb7075b953a8bb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### Inf-Retriever-V1 P@5: 21.1. Result ID: 69aa7a576dbb7075b953a8bd. [Reporting paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LLM"}. This record does not assert independent reproduction. ### BM25 P@5: 17.9. Result ID: 69aa7a576dbb7075b953a8b6. [Reporting paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"Classical"}. This record does not assert independent reproduction. ### PromptCase P@5: 17.1. Result ID: 69aa7a576dbb7075b953a8ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LM"}. This record does not assert independent reproduction. ### SAILER P@5: 16.6. Result ID: 69aa7a576dbb7075b953a8b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LM"}. This record does not assert independent reproduction. ### LEGAL-BERT P@5: 4.47. Result ID: 69aa7a576dbb7075b953a8b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LM"}. This record does not assert independent reproduction. ### MonoT5 P@5: 0.71. Result ID: 69aa7a576dbb7075b953a8b8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Stage":"One-stage","Architecture":"LM"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.