# Curiosity on G2U Curiosity Source: [Curiosity on G2U Curiosity]() Collection index: [llms.txt]() Data updated: 2026-05-15T07:24:23.371Z Evaluation of Curiosity performance on the G2U Curiosity dataset, reporting metrics such as Mean, Average Rank, and Ratio@Top-3. 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 | Curiosity | | Dataset (including any stated split) | G2U Curiosity | | Metric | Mean | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 30 of 30 | | 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 - [Mean](): Higher is better; unit: SCALAR. - [Average Rank](): Lower is better; unit: SCALAR. - [Ratio@Top-3 (%)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Mean | Result date | Report paper | | --- | --- | --- | --- | | V-JEPA2 | 1.402 | Not recorded | 691ea07fd6b0dd334292df3b | | GenieRedux | 1.374 | Not recorded | 691ea07fd6b0dd334292df3b | | GenieRedux | 1.248 | Not recorded | 691ea07fd6b0dd334292df3b | | GenieRedux | 1.198 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-8B | 1.173 | Not recorded | 691ea07fd6b0dd334292df3b | | DreamerV3 | 1.161 | Not recorded | 691ea07fd6b0dd334292df3b | | V-JEPA2 | 1.15 | Not recorded | 691ea07fd6b0dd334292df3b | | V-JEPA2 | 1.136 | Not recorded | 691ea07fd6b0dd334292df3b | | DreamerV3 | 1.12 | Not recorded | 691ea07fd6b0dd334292df3b | | Human | 1 | Not recorded | 691ea07fd6b0dd334292df3b | | DreamerV3 | 0.964 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-30B-A3B | 0.878 | Not recorded | 691ea07fd6b0dd334292df3b | | DQN | 0.856 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-8B-BC | 0.812 | Not recorded | 691ea07fd6b0dd334292df3b | | ACT-BC | 0.793 | Not recorded | 691ea07fd6b0dd334292df3b | | DQN | 0.772 | Not recorded | 691ea07fd6b0dd334292df3b | | DQN | 0.62 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-30B-A3B | 0.528 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-4o | 0.368 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-30B-A3B | 0.325 | Not recorded | 691ea07fd6b0dd334292df3b | | Qwen3-VL-8B | 0.325 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-4o | 0.319 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-5 | 0.298 | Not recorded | 691ea07fd6b0dd334292df3b | | PPO | 0.233 | Not recorded | 691ea07fd6b0dd334292df3b | | PPO | 0.22 | Not recorded | 691ea07fd6b0dd334292df3b | | PPO | 0.19 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-5 | 0.185 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-5 | 0.127 | Not recorded | 691ea07fd6b0dd334292df3b | | GPT-4o | 0.039 | Not recorded | 691ea07fd6b0dd334292df3b | | Random | 0 | Not recorded | 691ea07fd6b0dd334292df3b | ### V-JEPA2 Mean: 1.402. Result ID: 6a06ca2047f34ce924edffc2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### GenieRedux Mean: 1.374. Result ID: 6a06ca2047f34ce924edffc5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### GenieRedux Mean: 1.248. Result ID: 6a06ca2047f34ce924edffc6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### GenieRedux Mean: 1.198. Result ID: 6a06ca2047f34ce924edffc4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Mean: 1.173. Result ID: 6a06ca2047f34ce924edffd1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Interactive Physical Reasoner Group","Interactive Physical Reasoner (IPR) Status":"w/ IPR"}. This record does not assert independent reproduction. ### DreamerV3 Mean: 1.161. Result ID: 6a06ca2047f34ce924edffbf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### V-JEPA2 Mean: 1.15. Result ID: 6a06ca2047f34ce924edffc1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### V-JEPA2 Mean: 1.136. Result ID: 6a06ca2047f34ce924edffc3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### DreamerV3 Mean: 1.12. Result ID: 6a06ca2047f34ce924edffbe. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### Human Mean: 1. Result ID: 6a06ca2047f34ce924edffb5. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"Control Group"}. This record does not assert independent reproduction. ### DreamerV3 Mean: 0.964. Result ID: 6a06ca2047f34ce924edffc0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"World Model Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### Qwen3-VL-30B-A3B Mean: 0.878. Result ID: 6a06ca2047f34ce924edffce. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### DQN Mean: 0.856. Result ID: 6a06ca2047f34ce924edffbb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### Qwen3-VL-8B-BC Mean: 0.812. Result ID: 6a06ca2047f34ce924edffb7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Imitation Learning (IL) Group"}. This record does not assert independent reproduction. ### ACT-BC Mean: 0.793. Result ID: 6a06ca2047f34ce924edffb6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Imitation Learning (IL) Group"}. This record does not assert independent reproduction. ### DQN Mean: 0.772. Result ID: 6a06ca2047f34ce924edffbc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### DQN Mean: 0.62. Result ID: 6a06ca2047f34ce924edffbd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### Qwen3-VL-30B-A3B Mean: 0.528. Result ID: 6a06ca2047f34ce924edffcf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### GPT-4o Mean: 0.368. Result ID: 6a06ca2047f34ce924edffc8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### Qwen3-VL-30B-A3B Mean: 0.325. Result ID: 6a06ca2047f34ce924edffcd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Mean: 0.325. Result ID: 6a06ca2047f34ce924edffd0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Interactive Physical Reasoner Group","Interactive Physical Reasoner (IPR) Status":"w/o IPR"}. This record does not assert independent reproduction. ### GPT-4o Mean: 0.319. Result ID: 6a06ca2047f34ce924edffc9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### GPT-5 Mean: 0.298. Result ID: 6a06ca2047f34ce924edffcb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### PPO Mean: 0.233. Result ID: 6a06ca2047f34ce924edffb8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### PPO Mean: 0.22. Result ID: 6a06ca2047f34ce924edffba. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### PPO Mean: 0.19. Result ID: 6a06ca2047f34ce924edffb9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Reinforcement Learning (RL) Group","Optimization Objective (@)":"curiosity"}. This record does not assert independent reproduction. ### GPT-5 Mean: 0.185. Result ID: 6a06ca2047f34ce924edffcc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"utility"}. This record does not assert independent reproduction. ### GPT-5 Mean: 0.127. Result ID: 6a06ca2047f34ce924edffca. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### GPT-4o Mean: 0.039. Result ID: 6a06ca2047f34ce924edffc7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Paradigm":"Multimodal Large Language Model (MLLM) Group","Optimization Objective (@)":"survival"}. This record does not assert independent reproduction. ### Random Mean: 0. Result ID: 6a06ca2047f34ce924edffb4. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"Control Group"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.