daslab-testing/Apertus-8B-2509-microQAT-logits
收藏Hugging Face2026-04-07 更新2026-04-12 收录
下载链接：
https://hf-mirror.com/datasets/daslab-testing/Apertus-8B-2509-microQAT-logits
下载链接
链接失效反馈官方服务：
资源简介：
---
license: apache-2.0
task_categories:
- text-generation
language:
- fr
- de
- en
- rm
- it
---
This dataset provides a small sample of TOP-K logits computed using [swiss-ai/Apertus-8B-2509](https://huggingface.co/swiss-ai/Apertus-8B-2509) on samples from Data Phase 5 of Apertus pre-training.
### Format
This data represents documents packed into chuncks of 4096 tokens separated by EOS. The provided fields are as follows:
- `input_ids`: Input tokens.
- `index`: Positions of top-256 highest-probability next-token predictions for each token.
- `exp_logits`: Normalized probabilities of top-256 highest-probability next-token predictions for each token.
### Purpose
This data is designed to be used for [Quantization-Aware Distillation](https://research.nvidia.com/labs/nemotron/files/NVFP4-QAD-Report.pdf) and evaluation of base Apertus models.
The data is presented in two splits:
- `train`: To be used for distillation.
- `val`: To be used for NTP and KL-Div loss validation.
# Data Composition
For more details, refer to **Section 3.3** of the [Apertus report](https://arxiv.org/pdf/2509.14233#page=20.69).
提供机构：
daslab-testing
