# Confidence Estimation (Freeform Tagging) on PopQA

Source: [Confidence Estimation (Freeform Tagging) on PopQA](<https://www.sota2.com/research/sota/confidence-estimation-freeform-tagging-on-popqa>)

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

Data updated: 2026-05-15T07:01:08.189Z

Evaluation of model confidence estimation performance on the PopQA dataset using the Freeform Tagging methodology, measured using Brier Score (BS), Expected Calibration Error (ECE-M), and Scaled Calibration (SC).

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 | Confidence Estimation (Freeform Tagging) |
| Dataset (including any stated split) | PopQA |
| Metric | Brier Score (BS) |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 3 of 3 |
| 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

- [Brier Score (BS)](<https://www.sota2.com/research/sota/confidence-estimation-freeform-tagging-on-popqa.md?metric=Brier%20Score%20(BS)>): Lower is better; unit: PERCENT.

- [ECE-M](<https://www.sota2.com/research/sota/confidence-estimation-freeform-tagging-on-popqa.md?metric=ECE-M>): Lower is better; unit: PERCENT.

- [SC](<https://www.sota2.com/research/sota/confidence-estimation-freeform-tagging-on-popqa.md?metric=SC>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Brier Score (BS) | Result date | Report paper |
| --- | --- | --- | --- |
| LOVEC-GRPO | 10.1 | Not recorded | 68fa78e61e209131cea7b12b |
| LOVEC-DPO | 10.3 | Not recorded | 68fa78e61e209131cea7b12b |
| LOVEC-SFT | 19.4 | Not recorded | 68fa78e61e209131cea7b12b |

### LOVEC-GRPO

Brier Score (BS): 10.1. Result ID: 6a06c4b047f34ce924edd321.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

Recorded evaluation configuration: {"Tagging Format":"Freeform","Backbone":"Llama3-8B-Instruct","Decoding Strategy":"Greedy"}. This record does not assert independent reproduction.

### LOVEC-DPO

Brier Score (BS): 10.3. Result ID: 6a06c4b047f34ce924edd322.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

Recorded evaluation configuration: {"Tagging Format":"Freeform","Backbone":"Llama3-8B-Instruct","Decoding Strategy":"Greedy"}. This record does not assert independent reproduction.

### LOVEC-SFT

Brier Score (BS): 19.4. Result ID: 6a06c4b047f34ce924edd320.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78e61e209131cea7b12b.md>)

Recorded evaluation configuration: {"Tagging Format":"Freeform","Backbone":"Llama3-8B-Instruct","Decoding Strategy":"Greedy"}. This record does not assert independent reproduction.

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