# Psychological Defense Detection on PsyDefDetect (test) Source: [Psychological Defense Detection on PsyDefDetect (test)]() Collection index: [llms.txt]() Data updated: 2026-06-02T08:46:05.434Z Evaluation of models for Psychological Defense Detection using the test split of the PsyDefDetect dataset, reporting classification metrics including Accuracy, Precision, Recall, and F1 scores. 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 | Psychological Defense Detection | | Dataset (including any stated split) | PsyDefDetect (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Macro F1 Score](): Higher is better; unit: PERCENT. - [Macro Precision](): Higher is better; unit: PERCENT. - [Macro Recall](): Higher is better; unit: PERCENT. - [Macro F1 (1-8)](): Higher is better; unit: PERCENT. - [Precision](): Higher is better; unit: PERCENT. - [Recall](): Higher is better; unit: PERCENT. - [F1 Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Ministral-8B (Local) | 64.71 | Not recorded | 6a1e63c7dbd7150506f853d1 | | Qwen3-8B | 64.19 | Not recorded | 6a1e63c7dbd7150506f853d1 | | RoBERTa-base + OS (unif.) | 61.8 | Not recorded | 6a1e63c7dbd7150506f853d1 | | RoBERTa-base + OS (tgt.) | 61.4 | Not recorded | 6a1e63c7dbd7150506f853d1 | | DeBERTa-v3-base (5-fold) | 59.1 | Not recorded | 6a1e63c7dbd7150506f853d1 | | RoBERTa-base | 51.3 | Not recorded | 6a1e63c7dbd7150506f853d1 | | RoBERTa-base (len=320) | 47.9 | Not recorded | 6a1e63c7dbd7150506f853d1 | | Qwen3-8B | 34.22 | Not recorded | 6a1e63c7dbd7150506f853d1 | | Llama 3.1-8B | 30.36 | Not recorded | 6a1e63c7dbd7150506f853d1 | | Ministral-8B | 24.06 | Not recorded | 6a1e63c7dbd7150506f853d1 | | Mistral-7B-v0.3 (Kaggle) | 14.44 | Not recorded | 6a1e63c7dbd7150506f853d1 | ### Ministral-8B (Local) Accuracy: 64.71. Result ID: 6a1e97bffbc373592c18d1a4. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"LLM Fine-Tuning (LoRA)"}. This record does not assert independent reproduction. ### Qwen3-8B Accuracy: 64.19. Result ID: 6a1e97bffbc373592c18d1a5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"LLM Fine-Tuning (LoRA)"}. This record does not assert independent reproduction. ### RoBERTa-base + OS (unif.) Accuracy: 61.8. Result ID: 6a1e97bffbc373592c18d19e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"BERT-family (Supervised)","Input / Loss":"\[CTX\]/\[TGT\], Std CE + OS(120), LR 2e-5","Epochs":"6"}. This record does not assert independent reproduction. ### RoBERTa-base + OS (tgt.) Accuracy: 61.4. Result ID: 6a1e97bffbc373592c18d19f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"BERT-family (Supervised)","Input / Loss":"\[CTX\]/\[TGT\], Std CE + tgt. OS, LR 2e-5","Epochs":"8"}. This record does not assert independent reproduction. ### DeBERTa-v3-base (5-fold) Accuracy: 59.1. Result ID: 6a1e97bffbc373592c18d19b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"BERT-family (Supervised)","Input / Loss":"\[CTX\]/\[TGT\], Wtd CE, LR 1.5e-5","Epochs":"8"}. This record does not assert independent reproduction. ### RoBERTa-base Accuracy: 51.3. Result ID: 6a1e97bffbc373592c18d19c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"BERT-family (Supervised)","Input / Loss":"\[CTX\]/\[TGT\], Wtd CE, LR 2e-5","Epochs":"6"}. This record does not assert independent reproduction. ### RoBERTa-base (len=320) Accuracy: 47.9. Result ID: 6a1e97bffbc373592c18d19d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"BERT-family (Supervised)","Input / Loss":"\[CTX\]/\[TGT\], Wtd CE, LR 1.5e-5","Epochs":"6"}. This record does not assert independent reproduction. ### Qwen3-8B Accuracy: 34.22. Result ID: 6a1e97bffbc373592c18d1a0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Zero-Shot LLM Inference"}. This record does not assert independent reproduction. ### Llama 3.1-8B Accuracy: 30.36. Result ID: 6a1e97bffbc373592c18d1a1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Zero-Shot LLM Inference"}. This record does not assert independent reproduction. ### Ministral-8B Accuracy: 24.06. Result ID: 6a1e97bffbc373592c18d1a2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Zero-Shot LLM Inference"}. This record does not assert independent reproduction. ### Mistral-7B-v0.3 (Kaggle) Accuracy: 14.44. Result ID: 6a1e97bffbc373592c18d1a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"LLM Fine-Tuning (LoRA)"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.