llm-annotations
JQL-LLM-Edu-Annotations
📚 JQL Educational Quality Annotations from LLMs
This dataset provides 17,186,606 documents with high-quality LLM annotations for evaluating the educational value of web documents, and serves as a benchmark for training and evaluating multilingual LLM annotators as described in the JQL paper.
📝 Dataset Summary
Multilingual document-level quality annotations scored on a 0–5 educational value scale by three state-of-the-art LLMs:
Gemma-3-27B-it, Mistral-3.1-24B-it… See the full description on the dataset page: https://huggingface.co/datasets/JQL-AI/JQL-LLM-Edu-Annotations.narrative-llm-annotations
NarraDolma LLM-Labeled — Distillation Set
The intermediate, LLM-labeled dataset that bridges the small human gold set and the
full NarraDolma corpus. It contains 5,000 passages sampled from
Dolma and labeled by Gemma across
all 11 narrative dimensions, stratified by source and topic to preserve the original
distribution. These labels are the knowledge-distillation training set used to
train NarraBert.
Paper: arXiv:2606.19468
Collection: Narratives in LLM Pretraining Data… See the full description on the dataset page: https://huggingface.co/datasets/CLS-Lab/narrative-llm-annotations.llm-delusion-response-annotations
LLM Delusion-Like Belief Reinforcement Annotations
This dataset contains human annotations of responses generated by conversational large language models (LLMs) to prompts expressing potentially delusion-like or reality-distorted beliefs.
The purpose of the dataset is to support evaluation of whether conversational LLM responses may unintentionally reinforce or strengthen delusion-like beliefs.
Dataset Files
Consensus Dataset… See the full description on the dataset page: https://huggingface.co/datasets/ManjuKrish/llm-delusion-response-annotations.llm-delusion-response-annotations
LLM Delusion-Like Belief Reinforcement Annotations
This dataset contains human annotations of responses generated by conversational large language models (LLMs) to prompts expressing potentially delusion-like or reality-distorted beliefs.
The purpose of the dataset is to support evaluation of whether conversational LLM responses may unintentionally reinforce or strengthen delusion-like beliefs.
Dataset Files
Consensus Dataset… See the full description on the dataset page: https://huggingface.co/datasets/vennu95/llm-delusion-response-annotations.the_stack_llm_annotations
Dataset
이 데이터셋은 fineweb-edu의 방법을 여러 프로그래밍 언어에 적용하기 위해 만들어진 합성 데이터셋입니다.
기존에 존재하던 HuggingFaceTB/smollm-corpus의 Python-edu는 Python으로만 한정되어 있었습니다. 이 데이터셋은 bigcode/the-stack-dedup에서 21개의 프로그래밍 언어에서 각각 30k 샘플을 추출해 평가해 여러 언어에 대응합니다.
구체적으로는 devngho/the-stack-mini-nonshuffled의 첫 30k 샘플이 사용되었습니다.
This synthetic dataset was created to apply the methods of fineweb-edu to multiple programming language datasets.
The previously existing Python-edu of HuggingFaceTB/smollm-corpus was limited to… See the full description on the dataset page: https://huggingface.co/datasets/devngho/the_stack_llm_annotations.the-stack-llm-annotations-v2
Dataset
이 데이터셋은 fineweb-edu의 방법을 여러 프로그래밍 언어에 적용하기 위해 만들어진 합성 데이터셋입니다.
기존에 존재하던 HuggingFaceTB/smollm-corpus의 Python-edu는 Python으로만 한정되어 있었습니다. 이 데이터셋은 bigcode/the-stack-dedup에서 21개의 프로그래밍 언어에서 각각 30k 샘플을 추출해 평가해 여러 언어에 대응합니다.
구체적으로는 devngho/the-stack-mini-nonshuffled의 MIT, Apache 2.0, BSD 2-clause, BSD 3-clause 라이선스인 첫 30k 샘플이 사용되었습니다.
devngho/the_stack_llm_annotations와 유사하나, 평가에 Qwen2.5-32B 대신 Qwen2.5-Coder-32B를 사용했습니다.
This synthetic dataset was created to apply the methods of… See the full description on the dataset page: https://huggingface.co/datasets/devngho/the-stack-llm-annotations-v2.