---
license: apache-2.0
task_categories:
- text-generation
language:
- km
size_categories:
- 10K<n<100K
---

# Dataset Card for ```attentionlab/qa```

<!-- Provide a quick summary of the dataset. -->

QA Results and Dataset used to evaluate [attentionlab/bayon-it](https://huggingface.co/attentionlab/bayon-it), [attentionlab/bayon](https://huggingface.co/attentionlab/bayon), [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct), [openai/gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b), [aisingapore/Qwen-SEA-LION-v4.5-27B-IT](https://huggingface.co/aisingapore/Qwen-SEA-LION-v4.5-27B-IT), [google/gemma-4-31B-it](https://huggingface.co/google/gemma-4-31B-it), and [zai-org/GLM-4.7](https://huggingface.co/zai-org/GLM-4.7) for the AACL-IJCNLP 2026 SRW paper "Small Model, Native Design: Exploring Linguistic Purity and Reduced Language Bleeding in a 100M Khmer Language Model"