datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
magpie_llama70b_200k_filtered_swedish
Short description
Roughly 200k filtered instruction : response pairs in Swedish, filtered from roughtly 550k.
Contains "normal" QA along with math and coding QA.
Filtering, removed:
Deduplications
Instructions scored less than good or excellent
Responses scored less than -10 from ArmoRM-Llama3-8B-v0.1
Instructions and responses less than 10 in length or more than 2048
Usage
from datasets import load_dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/nicher92/magpie_llama70b_200k_filtered_swedish.magpie_llama70b_260k_filtered_swedish
Short description
Roughly 260k filtered instruction : response pairs in Swedish, filtered from roughtly 650k.
Contains "normal" QA along with math and coding QA and multiple choice questions and answers.
Filtering, removed:
Deduplications
Instructions scored less than good or excellent
Responses scored less than -10 from ArmoRM-Llama3-8B-v0.1
Instructions and responses less than 10 in length or more than 2048
Usage
from datasets import load_dataset
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/nicher92/magpie_llama70b_260k_filtered_swedish.swe-atlas-qna-dev41-mcode-m3
SWE-Atlas-QnA dev-41 — mcode + MiniMax-M3 trajectories
Complete run logs for one sweep of the 41-task SWE-Atlas-QnA dev slice, with the
LLM-rubric judge output for every task. Published as a reference trajectory set for
explore/comprehension evaluation.
41 of 41 tasks scored, mean agg_score 0.815 (median 0.875, min 0.222); mean
reward 0.244; 10 tasks (24 %) at the 1.0 ceiling.
suite
ai-solution-finetune/swe-atlas-qna-dev-50 minus 9 musl tasks = 41
agent
mcode… See the full description on the dataset page: https://huggingface.co/datasets/miaomiao64/swe-atlas-qna-dev41-mcode-m3.magpie_llama70b_100k_swedish
Dataset Card
Llama 3.3 generated instruction: response pairs using the MagPie pipeline: https://github.com/magpie-align/magpie/ https://arxiv.org/abs/2406.08464
Larger and filtered dataset available here:
nicher92/magpie_llama70b_200k_filtered_swedish
Dataset Details
Dataset Description
System prompt template for generating instructions: "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nDu är en hjälpsam AI… See the full description on the dataset page: https://huggingface.co/datasets/nicher92/magpie_llama70b_100k_swedish.