# Next Token Prediction on Amber 1.2T tokens

Source: [Next Token Prediction on Amber 1.2T tokens](<https://www.sota2.com/research/sota/next-token-prediction-on-amber-1-2t-tokens>)

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

Data updated: 2026-02-26T05:37:25.266Z

Evaluation of language model performance on the Amber 1.2T tokens dataset using Next Token Prediction, measured by BPD, Top-1 Error, and Top-100 Error.

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 | Next Token Prediction |
| Dataset (including any stated split) | Amber 1.2T tokens |
| Metric | BPD |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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

- [BPD](<https://www.sota2.com/research/sota/next-token-prediction-on-amber-1-2t-tokens.md?metric=BPD>): Lower is better; unit: SCALAR.

- [Top-1 Error](<https://www.sota2.com/research/sota/next-token-prediction-on-amber-1-2t-tokens.md?metric=Top-1%20Error>): Lower is better; unit: PERCENT.

- [Top-100 Error](<https://www.sota2.com/research/sota/next-token-prediction-on-amber-1-2t-tokens.md?metric=Top-100%20Error>): Lower is better; unit: PERCENT.

## Recorded results

| Method | BPD | Result date | Report paper |
| --- | --- | --- | --- |
| LLAMA2-7B | 4.28 | Not recorded | 68fa74401e209131cea6086f |
| LLAMA2-13B | 4.51 | Not recorded | 68fa74401e209131cea6086f |
| LLaMA2-70B | 6.39 | Not recorded | 68fa74401e209131cea6086f |
| Random Guess | 14.97 | Not recorded | 68fa74401e209131cea6086f |

### LLAMA2-7B

BPD: 4.28. Result ID: 699336ef3a6c3148491479e8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74401e209131cea6086f.md>)

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

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

Recorded evaluation configuration: {"quantization":"2-bit post-training QuIP"}. This record does not assert independent reproduction.

### LLAMA2-13B

BPD: 4.51. Result ID: 699336ef3a6c3148491479e9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74401e209131cea6086f.md>)

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

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

Recorded evaluation configuration: {"quantization":"2-bit post-training QuIP"}. This record does not assert independent reproduction.

### LLaMA2-70B

BPD: 6.39. Result ID: 699336ef3a6c3148491479ea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74401e209131cea6086f.md>)

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

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

Recorded evaluation configuration: {"quantization":"2-bit post-training QuIP"}. This record does not assert independent reproduction.

### Random Guess

BPD: 14.97. Result ID: 699336ef3a6c3148491479eb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74401e209131cea6086f.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.
