# Sycophancy Evaluation on DebateQA

Source: [Sycophancy Evaluation on DebateQA](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa>)

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

Data updated: 2026-04-06T07:36:33.993Z

Evaluation of model performance on the DebateQA dataset, measuring sycophancy tendencies across different conditions and difficulty levels.

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 | Sycophancy Evaluation |
| Dataset (including any stated split) | DebateQA |
| Metric | S (PD, L) |
| Unit | SCALAR |
| 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

- [S (PD, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PD%2C%20L)>): Higher is better; unit: SCALAR.

- [S (PD, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PD%2C%20M)>): Higher is better; unit: SCALAR.

- [S (PD, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PD%2C%20H)>): Higher is better; unit: SCALAR.

- [S (TD, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(TD%2C%20L)>): Higher is better; unit: SCALAR.

- [S (TD, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(TD%2C%20M)>): Higher is better; unit: SCALAR.

- [S (TD, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(TD%2C%20H)>): Higher is better; unit: SCALAR.

- [S (RI, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(RI%2C%20L)>): Higher is better; unit: SCALAR.

- [S (PI, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PI%2C%20L)>): Higher is better; unit: SCALAR.

- [S (RI, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(RI%2C%20M)>): Higher is better; unit: SCALAR.

- [S (PI, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PI%2C%20M)>): Higher is better; unit: SCALAR.

- [S (RI, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(RI%2C%20H)>): Higher is better; unit: SCALAR.

- [S (PI, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PI%2C%20H)>): Higher is better; unit: SCALAR.

- [S (NQ, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(NQ%2C%20L)>): Higher is better; unit: SCALAR.

- [S (PNQ, L)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PNQ%2C%20L)>): Higher is better; unit: SCALAR.

- [S (NQ, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(NQ%2C%20M)>): Higher is better; unit: SCALAR.

- [S (PNQ, M)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PNQ%2C%20M)>): Higher is better; unit: SCALAR.

- [S (NQ, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(NQ%2C%20H)>): Higher is better; unit: SCALAR.

- [S (PNQ, H)](<https://www.sota2.com/research/sota/sycophancy-evaluation-on-debateqa.md?metric=S%20(PNQ%2C%20H)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | S (PD, L) | Result date | Report paper |
| --- | --- | --- | --- |
| C. Opus | 0.481 | Not recorded | 69d33e107d0172140ec73704 |
| C. Sonnet | 0.428 | Not recorded | 69d33e107d0172140ec73704 |
| Gemma | 0.361 | Not recorded | 69d33e107d0172140ec73704 |
| Llama | 0.331 | Not recorded | 69d33e107d0172140ec73704 |
| Mistral | 0.228 | Not recorded | 69d33e107d0172140ec73704 |
| C. Haiku | 0.176 | Not recorded | 69d33e107d0172140ec73704 |

### C. Opus

S (PD, L): 0.481. Result ID: 69d3628047f34ce924badfba.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### C. Sonnet

S (PD, L): 0.428. Result ID: 69d3628047f34ce924badfb9.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Gemma

S (PD, L): 0.361. Result ID: 69d3628047f34ce924badfbc.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Llama

S (PD, L): 0.331. Result ID: 69d3628047f34ce924badfb7.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mistral

S (PD, L): 0.228. Result ID: 69d3628047f34ce924badfb8.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### C. Haiku

S (PD, L): 0.176. Result ID: 69d3628047f34ce924badfbb.

[Reporting paper](<https://www.sota2.com/research/paper/69d33e107d0172140ec73704.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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