# Generative multiple-choice on TruthfulQA (single)

Source: [Generative multiple-choice on TruthfulQA (single)](<https://www.sota2.com/research/sota/generative-multiple-choice-on-truthfulqa-single>)

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

Data updated: 2026-02-26T05:29:57.791Z

Evaluation of model performance on the specific 'single' configuration of the TruthfulQA dataset using a generative multiple-choice format, measured by Accuracy.

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 | Generative multiple-choice |
| Dataset (including any stated split) | TruthfulQA (single) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/generative-multiple-choice-on-truthfulqa-single.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| TACS-S (Sentence-level) | 78.1 | Not recorded | 68fa74ce1e209131cea63ff7 |
| TACS-T (Token-level) | 77.1 | Not recorded | 68fa74ce1e209131cea63ff7 |
| TACS-T (Token-level) | 62.5 | Not recorded | 68fa74ce1e209131cea63ff7 |
| TACS-S (Sentence-level) | 60.6 | Not recorded | 68fa74ce1e209131cea63ff7 |
| Mistral-Instruct-v0.2 | 54.7 | Not recorded | 68fa74ce1e209131cea63ff7 |
| Llama 2-Chat | 49.1 | Not recorded | 68fa74ce1e209131cea63ff7 |

### TACS-S (Sentence-level)

Accuracy: 78.1. Result ID: 69937c2e3a6c3148491646e2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Implementation](<https://github.com/ictnlp/TACS>)

Recorded evaluation configuration: {"Backbone":"Mistral-Instruct-v0.2","Model Size":"7B","Granularity":"Sentence-level","Information":"single"}. This record does not assert independent reproduction.

### TACS-T (Token-level)

Accuracy: 77.1. Result ID: 69937c2e3a6c3148491646e1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Implementation](<https://github.com/ictnlp/TACS>)

Recorded evaluation configuration: {"Backbone":"Mistral-Instruct-v0.2","Model Size":"7B","Granularity":"Token-level","Information":"single"}. This record does not assert independent reproduction.

### TACS-T (Token-level)

Accuracy: 62.5. Result ID: 69937c2e3a6c3148491646de.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Implementation](<https://github.com/ictnlp/TACS>)

Recorded evaluation configuration: {"Backbone":"Llama 2-Chat","Model Size":"7B","Granularity":"Token-level","Information":"single"}. This record does not assert independent reproduction.

### TACS-S (Sentence-level)

Accuracy: 60.6. Result ID: 69937c2e3a6c3148491646df.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Implementation](<https://github.com/ictnlp/TACS>)

Recorded evaluation configuration: {"Backbone":"Llama 2-Chat","Model Size":"7B","Granularity":"Sentence-level","Information":"single"}. This record does not assert independent reproduction.

### Mistral-Instruct-v0.2

Accuracy: 54.7. Result ID: 69937c2e3a6c3148491646e0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d61e209131cea4d909.md>)

[Implementation](<https://github.com/mistralai/mistral-src>)

Recorded evaluation configuration: {"Backbone":"Mistral-Instruct-v0.2","Model Size":"7B","Information":"single"}. This record does not assert independent reproduction.

### Llama 2-Chat

Accuracy: 49.1. Result ID: 69937c2e3a6c3148491646dd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ce1e209131cea63ff7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Backbone":"Llama 2-Chat","Model Size":"7B","Information":"single"}. This record does not assert independent reproduction.

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