# CarButton2 on Safety Gymnasium

Source: [CarButton2 on Safety Gymnasium](<https://www.sota2.com/research/sota/carbutton2-on-safety-gymnasium>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-30T07:42:41.097Z

Evaluation of agent performance on the CarButton2 task within the Safety Gymnasium environment, reporting Reward and Cost metrics.

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 | CarButton2 |
| Dataset (including any stated split) | Safety Gymnasium |
| Metric | Reward |
| Unit | SCALAR |
| 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

- [Reward](<https://www.sota2.com/research/sota/carbutton2-on-safety-gymnasium.md?metric=Reward>): Higher is better; unit: SCALAR.

- [Cost](<https://www.sota2.com/research/sota/carbutton2-on-safety-gymnasium.md?metric=Cost>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Reward | Result date | Report paper |
| --- | --- | --- | --- |
| CPQ | 0.37 | Not recorded | 69f2e23455a3c32af89fc7ec |
| CPQ | 0.37 | Not recorded | 69f2e23455a3c32af89fc7ec |
| CDT + SAS | 0.3 | Not recorded | 69f2e23455a3c32af89fc7ec |
| CDT + SAS | 0.3 | Not recorded | 69f2e23455a3c32af89fc7ec |
| DT + SAS | 0.14 | Not recorded | 69f2e23455a3c32af89fc7ec |
| DT + SAS | 0.14 | Not recorded | 69f2e23455a3c32af89fc7ec |
| CDT | 0.13 | Not recorded | 69f2e23455a3c32af89fc7ec |
| CDT | 0.13 | Not recorded | 69f2e23455a3c32af89fc7ec |
| DCRL | 0.09 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BCQ-Lag | 0.06 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BCQ-Lag | 0.06 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BC-Safe | -0.01 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BEAR-Lag | -0.01 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BC-Safe | -0.01 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BEAR-Lag | -0.01 | Not recorded | 69f2e23455a3c32af89fc7ec |
| COptiDICE | -0.07 | Not recorded | 69f2e23455a3c32af89fc7ec |
| COptiDICE | -0.07 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BC-All | -0.13 | Not recorded | 69f2e23455a3c32af89fc7ec |
| BC-All | -0.13 | Not recorded | 69f2e23455a3c32af89fc7ec |

### CPQ

Reward: 0.37. Result ID: 69f3071147f34ce924d923a6.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CPQ

Reward: 0.37. Result ID: 69f3071147f34ce924d9242f.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### CDT + SAS

Reward: 0.3. Result ID: 69f3071147f34ce924d923a0.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

Recorded evaluation configuration: {"Backbone":"CDT","Alignment strategy (SAS)":"Lyapunov-Guided"}. This record does not assert independent reproduction.

### CDT + SAS

Reward: 0.3. Result ID: 69f3071147f34ce924d92429.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### DT + SAS

Reward: 0.14. Result ID: 69f3071147f34ce924d9239f.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

Recorded evaluation configuration: {"Backbone":"DT","Alignment strategy (SAS)":"Lyapunov-Guided"}. This record does not assert independent reproduction.

### DT + SAS

Reward: 0.14. Result ID: 69f3071147f34ce924d92428.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### CDT

Reward: 0.13. Result ID: 69f3071147f34ce924d923a1.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: {"Backbone":"CDT"}. This record does not assert independent reproduction.

### CDT

Reward: 0.13. Result ID: 69f3071147f34ce924d9242a.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### DCRL

Reward: 0.09. Result ID: 69f3071147f34ce924d923a8.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26b7e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BCQ-Lag

Reward: 0.06. Result ID: 69f3071147f34ce924d923a4.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BCQ-Lag

Reward: 0.06. Result ID: 69f3071147f34ce924d9242d.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### BC-Safe

Reward: -0.01. Result ID: 69f3071147f34ce924d923a3.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BEAR-Lag

Reward: -0.01. Result ID: 69f3071147f34ce924d923a5.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BC-Safe

Reward: -0.01. Result ID: 69f3071147f34ce924d9242c.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### BEAR-Lag

Reward: -0.01. Result ID: 69f3071147f34ce924d9242e.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c381e209131cea32de5.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### COptiDICE

Reward: -0.07. Result ID: 69f3071147f34ce924d923a7.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c511e209131cea333c4.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### COptiDICE

Reward: -0.07. Result ID: 69f3071147f34ce924d92430.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c511e209131cea333c4.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

### BC-All

Reward: -0.13. Result ID: 69f3071147f34ce924d923a2.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BC-All

Reward: -0.13. Result ID: 69f3071147f34ce924d9242b.

[Reporting paper](<https://www.sota2.com/research/paper/69f2e23455a3c32af89fc7ec.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fc61e209131cea47551.md>)

Recorded evaluation configuration: {"cost thresholds":"3","evaluation episodes":"20","random seeds":"3"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
