# Question Answering Calibration on NQ, TQ, HQ, 2Wiki, Pararel

Source: [Question Answering Calibration on NQ, TQ, HQ, 2Wiki, Pararel](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel>)

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

Data updated: 2026-03-05T06:57:18.984Z

Evaluation of model calibration performance across a combined set of in-domain question answering datasets: NQ, TQ, HQ, 2Wiki, and Pararel.

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 | Question Answering Calibration |
| Dataset (including any stated split) | In-Domain Evaluation NQ, TQ, HQ, 2Wiki, Pararel |
| Metric | Calibration Error (NQ) |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 11 of 11 |
| 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

- [Calibration Error (NQ)](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Calibration%20Error%20(NQ)>): Lower is better; unit: SCALAR.

- [Calibration Error (TQ)](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Calibration%20Error%20(TQ)>): Lower is better; unit: SCALAR.

- [Calibration Error (HQ)](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Calibration%20Error%20(HQ)>): Lower is better; unit: SCALAR.

- [Calibration Error (2Wiki)](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Calibration%20Error%20(2Wiki)>): Lower is better; unit: SCALAR.

- [Calibration Error (Pararel)](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Calibration%20Error%20(Pararel)>): Lower is better; unit: SCALAR.

- [Average Calibration Error](<https://www.sota2.com/research/sota/question-answering-calibration-on-nq-tq-hq-2wiki-pararel.md?metric=Average%20Calibration%20Error>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Calibration Error (NQ) | Result date | Report paper |
| --- | --- | --- | --- |
| EliCal(1k) | 0.05 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Thermometer | 0.06 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Consis-Lex | 0.07 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| DACA | 0.09 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Elicitation | 0.09 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Cal-Only(1k) | 0.12 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Consis-Sem | 0.13 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Verbal-10 | 0.27 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Prob | 0.39 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| N-Prob | 0.39 | Not recorded | 68fa7c5b1e209131cea8dbc1 |
| Verbal-0 | 0.41 | Not recorded | 68fa7c5b1e209131cea8dbc1 |

### EliCal(1k)

Calibration Error (NQ): 0.05. Result ID: 69a928b36dbb7075b951f94c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","LoRA rank":"8","LoRA alpha":"16","Batch size":"16"}. This record does not assert independent reproduction.

### Thermometer

Calibration Error (NQ): 0.06. Result ID: 69a928b36dbb7075b951f948.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73d51e209131cea5e658.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### Consis-Lex

Calibration Error (NQ): 0.07. Result ID: 69a928b36dbb7075b951f946.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d71e209131cea4d943.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### DACA

Calibration Error (NQ): 0.09. Result ID: 69a928b36dbb7075b951f949.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c241e209131cea8ca97.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### Elicitation

Calibration Error (NQ): 0.09. Result ID: 69a928b36dbb7075b951f94a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","LoRA rank":"8","LoRA alpha":"16","Batch size":"16"}. This record does not assert independent reproduction.

### Cal-Only(1k)

Calibration Error (NQ): 0.12. Result ID: 69a928b36dbb7075b951f94b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67d84.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct","LoRA rank":"8","LoRA alpha":"16","Batch size":"16"}. This record does not assert independent reproduction.

### Consis-Sem

Calibration Error (NQ): 0.13. Result ID: 69a928b36dbb7075b951f947.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d71e209131cea4d943.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### Verbal-10

Calibration Error (NQ): 0.27. Result ID: 69a928b36dbb7075b951f945.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72941e209131cea573ff.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### Prob

Calibration Error (NQ): 0.39. Result ID: 69a928b36dbb7075b951f942.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff6d1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### N-Prob

Calibration Error (NQ): 0.39. Result ID: 69a928b36dbb7075b951f943.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff6d1.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

### Verbal-0

Calibration Error (NQ): 0.41. Result ID: 69a928b36dbb7075b951f944.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c5b1e209131cea8dbc1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72941e209131cea573ff.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction.

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