datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
spark2-5-tiny-fidelity-root-v1
spark2-5 random CPU fixture root
A root fidelity dataset in hidden form, produced by engines/tools/hf_capture.py from malaiwah/spark2-5-tiny-random-bf16.
The cut
the final hidden state handed to lm_head -- after the text model's final norm and immediately before the head matmul -- captured as the head module's input via torch.nn.Module.register_forward_pre_hook; replay applies the head ONLY (no final norm at replay time: the capture already sits after it). Same… See the full description on the dataset page: https://huggingface.co/datasets/malaiwah/spark2-5-tiny-fidelity-root-v1.arcee-ai__Llama-Spark-details
Dataset Card for Evaluation run of arcee-ai/Llama-Spark
Dataset automatically created during the evaluation run of model arcee-ai/Llama-Spark
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/arcee-ai__Llama-Spark-details.spark-capacity-boundary-study
Capacity-boundary optimization with Spark-X2.5-1.7B
This project contains an original evaluation for HER Hack-Astron #6. The report is in DISCUSSION.md. Creating or publishing these artifacts is not an award or payment.
The experiment checks whether increasing a six-item 0/1 knapsack's capacity by one causes the model to find the new optimum. Four seeded item families produce eight mathematical instances, each in English and Chinese. Every prompt runs once with thinking off and… See the full description on the dataset page: https://huggingface.co/datasets/aaaded/spark-capacity-boundary-study.mhpp
MHPP (questions only)
Evaluation-only questions for code/problem-solving. Do not train on test.
Data fields
id (int), question (string), prompt (string), function_name (string), parameters (list[str]), difficulty_types (int)
Jacoby746__Proto-Harpy-Spark-v0.1-7B-details
Dataset Card for Evaluation run of Jacoby746/Proto-Harpy-Spark-v0.1-7B
Dataset automatically created during the evaluation run of model Jacoby746/Proto-Harpy-Spark-v0.1-7B
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/Jacoby746__Proto-Harpy-Spark-v0.1-7B-details.sparkthai-model-arena
SparkThai Model Arena
Question bank, versioned snapshots, run history, and full per-trial results from the
SparkThai Model Arena — SparkThai's own
language-model eval suite, run on our own NVIDIA DGX Spark hardware.
🌐 sparkthai.com · 🏆 Model Arena · 🐙 github.com/sparkthai
What this is
The Model Arena grades Qwen model variants we host ourselves on DGX Spark across four kinds of checks:
code_exact — code tasks graded by running the model's code against an… See the full description on the dataset page: https://huggingface.co/datasets/sparkthai/sparkthai-model-arena.arcee-ai__Arcee-Spark-details
Dataset Card for Evaluation run of arcee-ai/Arcee-Spark
Dataset automatically created during the evaluation run of model arcee-ai/Arcee-Spark
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/arcee-ai__Arcee-Spark-details.