# Uncertainty Quantification on MAQA ∆K−1

Source: [Uncertainty Quantification on MAQA ∆K−1](<https://www.sota2.com/research/sota/uncertainty-quantification-on-maqa-k-1>)

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

Data updated: 2026-02-26T05:47:45.333Z

Evaluation of model performance on the MAQA ∆K−1 dataset specifically focusing on Uncertainty Quantification, measured using metrics like KL Divergence AUC.

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 | Uncertainty Quantification |
| Dataset (including any stated split) | MAQA ∆K−1 |
| Metric | KL Divergence AUC |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 28 of 28 |
| 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

- [KL Divergence AUC](<https://www.sota2.com/research/sota/uncertainty-quantification-on-maqa-k-1.md?metric=KL%20Divergence%20AUC>): Higher is better; unit: SCALAR.

## Recorded results

| Method | KL Divergence AUC | Result date | Report paper |
| --- | --- | --- | --- |
| Structure-Aware Minimum Bayes Risk Decoding | 0.757 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| Structure-Aware Minimum Bayes Risk Decoding | 0.756 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| Structure-Aware Minimum Bayes Risk Decoding | 0.743 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SE | 0.723 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| COCOA | 0.721 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SE | 0.72 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SAR | 0.718 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SE | 0.718 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SAR | 0.716 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SAR | 0.715 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| KLE | 0.708 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| KLE | 0.708 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| KLE | 0.705 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| COCOA | 0.704 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| COCOA | 0.699 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| MSP | 0.69 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| MSP | 0.689 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| MSP | 0.678 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| Structure-Aware Minimum Bayes Risk Decoding | 0.666 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SAR | 0.662 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| COCOA | 0.658 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| MSP | 0.655 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| KLE | 0.639 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| SE | 0.639 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| P(True) | 0.541 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| P(True) | 0.54 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| P(True) | 0.534 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |
| P(True) | 0.474 | Not recorded | 697c3b0ba8b01e76cc8da4c7 |

### Structure-Aware Minimum Bayes Risk Decoding

KL Divergence AUC: 0.757. Result ID: 6998fcd25698665ec4620c08.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### Structure-Aware Minimum Bayes Risk Decoding

KL Divergence AUC: 0.756. Result ID: 6998fcd25698665ec4620c0f.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### Structure-Aware Minimum Bayes Risk Decoding

KL Divergence AUC: 0.743. Result ID: 6998fcd25698665ec4620c1d.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### SE

KL Divergence AUC: 0.723. Result ID: 6998fcd25698665ec4620c06.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f141e209131cea43bf3.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### COCOA

KL Divergence AUC: 0.721. Result ID: 6998fcd25698665ec4620c19.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e91e209131cea7b2a8.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### SE

KL Divergence AUC: 0.72. Result ID: 6998fcd25698665ec4620c1b.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f141e209131cea43bf3.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### SAR

KL Divergence AUC: 0.718. Result ID: 6998fcd25698665ec4620c03.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73681e209131cea5c472.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### SE

KL Divergence AUC: 0.718. Result ID: 6998fcd25698665ec4620c0d.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f141e209131cea43bf3.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### SAR

KL Divergence AUC: 0.716. Result ID: 6998fcd25698665ec4620c18.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73681e209131cea5c472.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### SAR

KL Divergence AUC: 0.715. Result ID: 6998fcd25698665ec4620c0a.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73681e209131cea5c472.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### KLE

KL Divergence AUC: 0.708. Result ID: 6998fcd25698665ec4620c0c.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73601e209131cea5c05e.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### KLE

KL Divergence AUC: 0.708. Result ID: 6998fcd25698665ec4620c1a.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73601e209131cea5c05e.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### KLE

KL Divergence AUC: 0.705. Result ID: 6998fcd25698665ec4620c05.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73601e209131cea5c05e.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### COCOA

KL Divergence AUC: 0.704. Result ID: 6998fcd25698665ec4620c04.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e91e209131cea7b2a8.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### COCOA

KL Divergence AUC: 0.699. Result ID: 6998fcd25698665ec4620c0b.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e91e209131cea7b2a8.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### MSP

KL Divergence AUC: 0.69. Result ID: 6998fcd25698665ec4620c1c.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b71e209131cea7f100.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### MSP

KL Divergence AUC: 0.689. Result ID: 6998fcd25698665ec4620c07.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b71e209131cea7f100.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

### MSP

KL Divergence AUC: 0.678. Result ID: 6998fcd25698665ec4620c0e.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b71e209131cea7f100.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### Structure-Aware Minimum Bayes Risk Decoding

KL Divergence AUC: 0.666. Result ID: 6998fcd25698665ec4620c16.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### SAR

KL Divergence AUC: 0.662. Result ID: 6998fcd25698665ec4620c11.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73681e209131cea5c472.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### COCOA

KL Divergence AUC: 0.658. Result ID: 6998fcd25698665ec4620c12.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e91e209131cea7b2a8.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### MSP

KL Divergence AUC: 0.655. Result ID: 6998fcd25698665ec4620c15.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b71e209131cea7f100.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### KLE

KL Divergence AUC: 0.639. Result ID: 6998fcd25698665ec4620c13.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73601e209131cea5c05e.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### SE

KL Divergence AUC: 0.639. Result ID: 6998fcd25698665ec4620c14.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f141e209131cea43bf3.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### P(True)

KL Divergence AUC: 0.541. Result ID: 6998fcd25698665ec4620c17.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6df71e209131cea3d034.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-30B-A3B"}. This record does not assert independent reproduction.

### P(True)

KL Divergence AUC: 0.54. Result ID: 6998fcd25698665ec4620c10.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6df71e209131cea3d034.md>)

Recorded evaluation configuration: {"Backbone Model":"Qwen3-4B"}. This record does not assert independent reproduction.

### P(True)

KL Divergence AUC: 0.534. Result ID: 6998fcd25698665ec4620c09.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6df71e209131cea3d034.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-12B"}. This record does not assert independent reproduction.

### P(True)

KL Divergence AUC: 0.474. Result ID: 6998fcd25698665ec4620c02.

[Reporting paper](<https://www.sota2.com/research/paper/697c3b0ba8b01e76cc8da4c7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6df71e209131cea3d034.md>)

Recorded evaluation configuration: {"Backbone Model":"Gemma-3-4B"}. This record does not assert independent reproduction.

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