# Skill-selection on ToolBench-60 soft-0.3 preference regime (held-out pool)

Source: [Skill-selection on ToolBench-60 soft-0.3 preference regime (held-out pool)](<https://www.sota2.com/research/sota/skill-selection-on-toolbench-60-soft-0-3-preference-regime-held-out-pool>)

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

Data updated: 2026-06-05T09:11:26.697Z

Evaluation of the model's ability to select the correct skill using the ToolBench dataset under the soft-0.3 preference regime, measured by Regret, Accuracy, and SRC.

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 | Skill-selection |
| Dataset (including any stated split) | ToolBench-60 soft-0.3 preference regime (held-out pool) |
| Metric | Regret |
| Unit | SCALAR |
| Direction | Lower 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

- [Regret](<https://www.sota2.com/research/sota/skill-selection-on-toolbench-60-soft-0-3-preference-regime-held-out-pool.md?metric=Regret>): Lower is better; unit: SCALAR.

- [Accuracy](<https://www.sota2.com/research/sota/skill-selection-on-toolbench-60-soft-0-3-preference-regime-held-out-pool.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [SRC](<https://www.sota2.com/research/sota/skill-selection-on-toolbench-60-soft-0-3-preference-regime-held-out-pool.md?metric=SRC>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Regret | Result date | Report paper |
| --- | --- | --- | --- |
| Bandit-as-Override | 255.9 | Not recorded | 6a2258a186173ad9c1468be8 |
| Bandit-as-Override | 256.8 | Not recorded | 6a2258a186173ad9c1468be8 |
| Pure-Bandit | 279.2 | Not recorded | 6a2258a186173ad9c1468be8 |
| Pure-Bandit | 280.7 | Not recorded | 6a2258a186173ad9c1468be8 |
| Freq-as-Override | 286.5 | Not recorded | 6a2258a186173ad9c1468be8 |
| Freq-as-Override | 288 | Not recorded | 6a2258a186173ad9c1468be8 |
| Bandit-as-Context | 309.6 | Not recorded | 6a2258a186173ad9c1468be8 |
| Profile-Memory | 316 | Not recorded | 6a2258a186173ad9c1468be8 |
| Bandit-as-Context | 327.1 | Not recorded | 6a2258a186173ad9c1468be8 |
| Freq-Greedy | 327.2 | Not recorded | 6a2258a186173ad9c1468be8 |
| Profile-Memory | 327.4 | Not recorded | 6a2258a186173ad9c1468be8 |
| Freq-Greedy | 328 | Not recorded | 6a2258a186173ad9c1468be8 |
| InCotext-Memory | 372.5 | Not recorded | 6a2258a186173ad9c1468be8 |
| InCotext-Memory | 373.5 | Not recorded | 6a2258a186173ad9c1468be8 |
| ZeroShot-LLM | 375.5 | Not recorded | 6a2258a186173ad9c1468be8 |
| ZeroShot-LLM | 376.7 | Not recorded | 6a2258a186173ad9c1468be8 |
| Random | 491.5 | Not recorded | 6a2258a186173ad9c1468be8 |
| Random | 491.7 | Not recorded | 6a2258a186173ad9c1468be8 |

### Bandit-as-Override

Regret: 255.9. Result ID: 6a2291ddfbc373592c20dfb0.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"LLM with Statistical Prior"}. This record does not assert independent reproduction.

### Bandit-as-Override

Regret: 256.8. Result ID: 6a2291ddfbc373592c20dfd4.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Pure-Bandit

Regret: 279.2. Result ID: 6a2291ddfbc373592c20dfcf.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa60891e209131ce9efe66.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Pure-Bandit

Regret: 280.7. Result ID: 6a2291ddfbc373592c20dfab.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa60891e209131ce9efe66.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"Statistical"}. This record does not assert independent reproduction.

### Freq-as-Override

Regret: 286.5. Result ID: 6a2291ddfbc373592c20dfd3.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Freq-as-Override

Regret: 288. Result ID: 6a2291ddfbc373592c20dfaf.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"LLM with Statistical Prior"}. This record does not assert independent reproduction.

### Bandit-as-Context

Regret: 309.6. Result ID: 6a2291ddfbc373592c20dfae.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa60891e209131ce9efe66.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"LLM with Statistical Prior"}. This record does not assert independent reproduction.

### Profile-Memory

Regret: 316. Result ID: 6a2291ddfbc373592c20dfd1.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Bandit-as-Context

Regret: 327.1. Result ID: 6a2291ddfbc373592c20dfd2.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa60891e209131ce9efe66.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Freq-Greedy

Regret: 327.2. Result ID: 6a2291ddfbc373592c20dfce.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Profile-Memory

Regret: 327.4. Result ID: 6a2291ddfbc373592c20dfad.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"LLM with Memory"}. This record does not assert independent reproduction.

### Freq-Greedy

Regret: 328. Result ID: 6a2291ddfbc373592c20dfaa.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"Statistical"}. This record does not assert independent reproduction.

### InCotext-Memory

Regret: 372.5. Result ID: 6a2291ddfbc373592c20dfac.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"LLM with Memory"}. This record does not assert independent reproduction.

### InCotext-Memory

Regret: 373.5. Result ID: 6a2291ddfbc373592c20dfd0.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### ZeroShot-LLM

Regret: 375.5. Result ID: 6a2291ddfbc373592c20dfcd.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### ZeroShot-LLM

Regret: 376.7. Result ID: 6a2291ddfbc373592c20dfa9.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"No Learning"}. This record does not assert independent reproduction.

### Random

Regret: 491.5. Result ID: 6a2291ddfbc373592c20dfcc.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"GPT-5.2"}. This record does not assert independent reproduction.

### Random

Regret: 491.7. Result ID: 6a2291ddfbc373592c20dfa8.

[Reporting paper](<https://www.sota2.com/research/paper/6a2258a186173ad9c1468be8.md>)

Recorded evaluation configuration: {"Backbone":"DeepSeek-V4-Flash","Agent Category":"No Learning"}. This record does not assert independent reproduction.

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