# Structured Occlusion Reasoning on nuScenes PKL-guided v1.0 (val) Source: [Structured Occlusion Reasoning on nuScenes PKL-guided v1.0 (val)]() Collection index: [llms.txt]() Data updated: 2026-07-02T07:36:01.251Z Evaluation of Structured Occlusion Reasoning performance on the validation split of the nuScenes PKL-guided v1.0 dataset, covering agent behavior, visual cues, and predicted actions. 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 | Structured Occlusion Reasoning | | Dataset (including any stated split) | nuScenes PKL-guided v1.0 (val) | | Metric | Agent Class Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [Agent Class Accuracy](): Higher is better; unit: PERCENT. - [Agent Class F1](): Higher is better; unit: PERCENT. - [Camera View Accuracy](): Higher is better; unit: PERCENT. - [Camera View F1](): Higher is better; unit: PERCENT. - [Visual Cue Accuracy](): Higher is better; unit: PERCENT. - [Visual Cue F1](): Higher is better; unit: PERCENT. - [Occlusion Type Accuracy](): Higher is better; unit: PERCENT. - [Occlusion Type F1](): Higher is better; unit: PERCENT. - [Occluder Class Accuracy](): Higher is better; unit: PERCENT. - [Occluder Class F1](): Higher is better; unit: PERCENT. - [Priority Accuracy](): Higher is better; unit: PERCENT. - [Priority F1](): Higher is better; unit: PERCENT. - [Sector Accuracy](): Higher is better; unit: PERCENT. - [Sector F1](): Higher is better; unit: PERCENT. - [Action BLEU](): Higher is better; unit: PERCENT. - [Action ROUGE](): Higher is better; unit: PERCENT. - [Action Similarity](): Higher is better; unit: PERCENT. - [Overall Average Accuracy](): Higher is better; unit: PERCENT. - [Overall Average F1](): Higher is better; unit: PERCENT. - [Overall Average Action Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Agent Class Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Intern-GPT | 70.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-4B | 64.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-1B | 64.4 | Not recorded | 6a45f083dc71372d7795a1d1 | | Qwen-72B | 51.6 | Not recorded | 6a45f083dc71372d7795a1d1 | | Qwen-32B | 51.2 | Not recorded | 6a45f083dc71372d7795a1d1 | | Gemma-4B | 50.4 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-8B | 49.6 | Not recorded | 6a45f083dc71372d7795a1d1 | | Gemma-27B | 48.4 | Not recorded | 6a45f083dc71372d7795a1d1 | | MiniCPM | 46.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | Qwen-7B | 43.2 | Not recorded | 6a45f083dc71372d7795a1d1 | | Qwen-3B | 32.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-14B | 32.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-2B | 32.4 | Not recorded | 6a45f083dc71372d7795a1d1 | | Gemma-12B | 30 | Not recorded | 6a45f083dc71372d7795a1d1 | | Intern-38B | 22 | Not recorded | 6a45f083dc71372d7795a1d1 | | DriveLM | 14.8 | Not recorded | 6a45f083dc71372d7795a1d1 | | DriveMM | 0.4 | Not recorded | 6a45f083dc71372d7795a1d1 | | DriveGPT4 | 0 | Not recorded | 6a45f083dc71372d7795a1d1 | ### Intern-GPT Agent Class Accuracy: 70.8. Result ID: 6a4614c360b88bb339842ff1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-4B Agent Class Accuracy: 64.8. Result ID: 6a4614c360b88bb339842fee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-1B Agent Class Accuracy: 64.4. Result ID: 6a4614c360b88bb339842fec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Qwen-72B Agent Class Accuracy: 51.6. Result ID: 6a4614c360b88bb339842feb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Qwen-32B Agent Class Accuracy: 51.2. Result ID: 6a4614c360b88bb339842fea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Gemma-4B Agent Class Accuracy: 50.4. Result ID: 6a4614c360b88bb339842ff3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-8B Agent Class Accuracy: 49.6. Result ID: 6a4614c360b88bb339842fef. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Gemma-27B Agent Class Accuracy: 48.4. Result ID: 6a4614c360b88bb339842ff5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### MiniCPM Agent Class Accuracy: 46.8. Result ID: 6a4614c360b88bb339842ff6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Qwen-7B Agent Class Accuracy: 43.2. Result ID: 6a4614c360b88bb339842fe9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Qwen-3B Agent Class Accuracy: 32.8. Result ID: 6a4614c360b88bb339842fe8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-14B Agent Class Accuracy: 32.8. Result ID: 6a4614c360b88bb339842ff0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-2B Agent Class Accuracy: 32.4. Result ID: 6a4614c360b88bb339842fed. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Gemma-12B Agent Class Accuracy: 30. Result ID: 6a4614c360b88bb339842ff4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### Intern-38B Agent Class Accuracy: 22. Result ID: 6a4614c360b88bb339842ff2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### DriveLM Agent Class Accuracy: 14.8. Result ID: 6a4614c360b88bb339842ff8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### DriveMM Agent Class Accuracy: 0.4. Result ID: 6a4614c360b88bb339842ff7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. ### DriveGPT4 Agent Class Accuracy: 0. Result ID: 6a4614c360b88bb339842ff9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.