# Statistical Robustness under Random Bit-Flip Attacks on Qwen 0.5B 2.5

Source: [Statistical Robustness under Random Bit-Flip Attacks on Qwen 0.5B 2.5](<https://www.sota2.com/research/sota/statistical-robustness-under-random-bit-flip-attacks-on-qwen-0-5b-2-5>)

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

Data updated: 2026-03-18T08:39:30.154Z

Evaluation of the statistical robustness of the Qwen 0.5B 2.5 model when subjected to random bit-flip attacks, measured by Mean Perplexity (PPL), Maximum PPL, and Failure Rate.

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 | Statistical Robustness under Random Bit-Flip Attacks |
| Dataset (including any stated split) | Qwen 0.5B 2.5 |
| Metric | Mean Perplexity (PPL) |
| Unit | SCALAR |
| 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

- [Mean Perplexity (PPL)](<https://www.sota2.com/research/sota/statistical-robustness-under-random-bit-flip-attacks-on-qwen-0-5b-2-5.md?metric=Mean%20Perplexity%20(PPL)>): Lower is better; unit: SCALAR.

- [Maximum Perplexity (PPL)](<https://www.sota2.com/research/sota/statistical-robustness-under-random-bit-flip-attacks-on-qwen-0-5b-2-5.md?metric=Maximum%20Perplexity%20(PPL)>): Lower is better; unit: SCALAR.

- [Failure Rate](<https://www.sota2.com/research/sota/statistical-robustness-under-random-bit-flip-attacks-on-qwen-0-5b-2-5.md?metric=Failure%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Mean Perplexity (PPL) | Result date | Report paper |
| --- | --- | --- | --- |
| RoR | 13.39 | Not recorded | 69ba31870636e5bf4afc07b5 |
| FaR | 13.98 | Not recorded | 69ba31870636e5bf4afc07b5 |
| Baseline | 14.24 | Not recorded | 69ba31870636e5bf4afc07b5 |

### RoR

Mean Perplexity (PPL): 13.39. Result ID: 69ba64a447f34ce924a457b3.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31870636e5bf4afc07b5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ba31870636e5bf4afc07b5.md>)

Recorded evaluation configuration: {"Number of Trials":"2,000"}. This record does not assert independent reproduction.

### FaR

Mean Perplexity (PPL): 13.98. Result ID: 69ba64a447f34ce924a457b2.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31870636e5bf4afc07b5.md>)

Recorded evaluation configuration: {"Number of Trials":"2,000"}. This record does not assert independent reproduction.

### Baseline

Mean Perplexity (PPL): 14.24. Result ID: 69ba64a447f34ce924a457b1.

[Reporting paper](<https://www.sota2.com/research/paper/69ba31870636e5bf4afc07b5.md>)

Recorded evaluation configuration: {"Number of Trials":"2,000"}. This record does not assert independent reproduction.

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