# Self-Harm Detection on JiraiBench 1.0 (test) Source: [Self-Harm Detection on JiraiBench 1.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:51:27.817Z Evaluation of model performance on the test split of the JiraiBench 1.0 dataset for the task of Self-Harm Detection, reporting Macro F1 score. 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 | Self-Harm Detection | | Dataset (including any stated split) | JiraiBench 1.0 (test) | | Metric | Macro F1 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 35 of 35 | | 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 - [Macro F1](): Higher is better; unit: PERCENT. ## Recorded results | Method | Macro F1 | Result date | Report paper | | --- | --- | --- | --- | | SAS (English version) | 66.63 | Not recorded | 69608b7653fdd5c3617443a2 | | S^3 Agent | 61.8 | Not recorded | 69608b7653fdd5c3617443a2 | | OWL | 59.75 | Not recorded | 69608b7653fdd5c3617443a2 | | Self-Refine | 57.92 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot | 57.33 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot CoT | 57.02 | Not recorded | 69608b7653fdd5c3617443a2 | | Plan-and-Solve | 55.91 | Not recorded | 69608b7653fdd5c3617443a2 | | SAS (English version) | 53.89 | Not recorded | 69608b7653fdd5c3617443a2 | | Self-Refine | 53.33 | Not recorded | 69608b7653fdd5c3617443a2 | | SAS (English version) | 50.32 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot CoT | 49.99 | Not recorded | 69608b7653fdd5c3617443a2 | | Plan-and-Solve | 49.92 | Not recorded | 69608b7653fdd5c3617443a2 | | SAS (English version) | 49.38 | Not recorded | 69608b7653fdd5c3617443a2 | | Self-Refine | 49.3 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot CoT | 48.86 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot | 48.33 | Not recorded | 69608b7653fdd5c3617443a2 | | Plan-and-Solve | 47.77 | Not recorded | 69608b7653fdd5c3617443a2 | | S^3 Agent | 47.18 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot CoT | 47.05 | Not recorded | 69608b7653fdd5c3617443a2 | | OWL | 47.03 | Not recorded | 69608b7653fdd5c3617443a2 | | OWL | 46.57 | Not recorded | 69608b7653fdd5c3617443a2 | | S^3 Agent | 46.03 | Not recorded | 69608b7653fdd5c3617443a2 | | Plan-and-Solve | 45.94 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot | 45.92 | Not recorded | 69608b7653fdd5c3617443a2 | | S^3 Agent | 45.82 | Not recorded | 69608b7653fdd5c3617443a2 | | SAS (English version) | 44.48 | Not recorded | 69608b7653fdd5c3617443a2 | | OWL | 43.12 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot | 39.71 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot | 39 | Not recorded | 69608b7653fdd5c3617443a2 | | Self-Refine | 38.83 | Not recorded | 69608b7653fdd5c3617443a2 | | Zero-shot CoT | 36.29 | Not recorded | 69608b7653fdd5c3617443a2 | | OWL | 35.83 | Not recorded | 69608b7653fdd5c3617443a2 | | Self-Refine | 34.8 | Not recorded | 69608b7653fdd5c3617443a2 | | S^3 Agent | 34.72 | Not recorded | 69608b7653fdd5c3617443a2 | | Plan-and-Solve | 32.74 | Not recorded | 69608b7653fdd5c3617443a2 | ### SAS (English version) Macro F1: 66.63. Result ID: 699696705698665ec4dfe5c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### S^3 Agent Macro F1: 61.8. Result ID: 699696705698665ec4dfe5b8. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### OWL Macro F1: 59.75. Result ID: 699696705698665ec4dfe5bd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### Self-Refine Macro F1: 57.92. Result ID: 699696705698665ec4dfe5b3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### Zero-shot Macro F1: 57.33. Result ID: 699696705698665ec4dfe5a4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### Zero-shot CoT Macro F1: 57.02. Result ID: 699696705698665ec4dfe5a9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### Plan-and-Solve Macro F1: 55.91. Result ID: 699696705698665ec4dfe5ae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"DeepSeek V3.2"}. This record does not assert independent reproduction. ### SAS (English version) Macro F1: 53.89. Result ID: 699696705698665ec4dfe5c3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### Self-Refine Macro F1: 53.33. Result ID: 699696705698665ec4dfe5b5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### SAS (English version) Macro F1: 50.32. Result ID: 699696705698665ec4dfe5c0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### Zero-shot CoT Macro F1: 49.99. Result ID: 699696705698665ec4dfe5a7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### Plan-and-Solve Macro F1: 49.92. Result ID: 699696705698665ec4dfe5ac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### SAS (English version) Macro F1: 49.38. Result ID: 699696705698665ec4dfe5c4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### Self-Refine Macro F1: 49.3. Result ID: 699696705698665ec4dfe5b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### Zero-shot CoT Macro F1: 48.86. Result ID: 699696705698665ec4dfe5aa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### Zero-shot Macro F1: 48.33. Result ID: 699696705698665ec4dfe5a5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### Plan-and-Solve Macro F1: 47.77. Result ID: 699696705698665ec4dfe5af. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### S^3 Agent Macro F1: 47.18. Result ID: 699696705698665ec4dfe5b6. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### Zero-shot CoT Macro F1: 47.05. Result ID: 699696705698665ec4dfe5ab. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### OWL Macro F1: 47.03. Result ID: 699696705698665ec4dfe5bb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### OWL Macro F1: 46.57. Result ID: 699696705698665ec4dfe5be. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### S^3 Agent Macro F1: 46.03. Result ID: 699696705698665ec4dfe5ba. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### Plan-and-Solve Macro F1: 45.94. Result ID: 699696705698665ec4dfe5b0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### Zero-shot Macro F1: 45.92. Result ID: 699696705698665ec4dfe5a6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### S^3 Agent Macro F1: 45.82. Result ID: 699696705698665ec4dfe5b9. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Ministral-3-8B"}. This record does not assert independent reproduction. ### SAS (English version) Macro F1: 44.48. Result ID: 699696705698665ec4dfe5c1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### OWL Macro F1: 43.12. Result ID: 699696705698665ec4dfe5bf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Gemma-3-12B-it"}. This record does not assert independent reproduction. ### Zero-shot Macro F1: 39.71. Result ID: 699696705698665ec4dfe5a2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### Zero-shot Macro F1: 39. Result ID: 699696705698665ec4dfe5a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### Self-Refine Macro F1: 38.83. Result ID: 699696705698665ec4dfe5b2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### Zero-shot CoT Macro F1: 36.29. Result ID: 699696705698665ec4dfe5a8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### OWL Macro F1: 35.83. Result ID: 699696705698665ec4dfe5bc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### Self-Refine Macro F1: 34.8. Result ID: 699696705698665ec4dfe5b1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Qwen-2.5-7B"}. This record does not assert independent reproduction. ### S^3 Agent Macro F1: 34.72. Result ID: 699696705698665ec4dfe5b7. [Reporting paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. ### Plan-and-Solve Macro F1: 32.74. Result ID: 699696705698665ec4dfe5ad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"LLM Backbone":"Llama-3.1-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.