datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
classified_images_gemmagemma4-serving-bench-data
Gemma 4 12B (QAT-Q4_0) — Serving-Behavior Test Data
Test data, charts, and the running research log from an autonomous research
loop characterizing and tuning a Gemma 4 12B QAT-Q4_0 model served via
llama.cpp/llamafile on a single RTX 3080 Ti. Every ~30 min the loop
summarizes findings, proposes a goal, tests it end-to-end, documents success or
failure, and publishes here + to GitHub.
Model under test: gemma-4-12b-it-qat-q4_0.gguf (Google, June 2026), 128K
ctx, f16 KV, MTP… See the full description on the dataset page: https://huggingface.co/datasets/SEBK4C/gemma4-serving-bench-data.gemini-3-pro-previewgemma-4-e2b-atlas
test_follow_benchGEMRec-PromptBook
GEMRec-18k -- Prompt Book
This is the official image dataset for the paper Towards Personalized Prompt-Model Retrieval for Generative Recommendation.
Dataset Intro
GEMRec-18K is a prompt-model interaction dataset with 18K images generated by 200 publicly-available generative models paired with a diverse set of 90 textual prompts. We randomly sampled a subset of 197 models from the full set of models (all finetuned from Stable Diffusion) on Civitai according to the… See the full description on the dataset page: https://huggingface.co/datasets/MAPS-research/GEMRec-PromptBook.rlhf-gemma3-indfood-1kMathCanvas_gemini3flash_seed42_idx_900_1050gemma-4-e4b-it-atlas
juiceb0xc0de/gemma-4-e4b-it-atlas
A brain atlas for google/gemma-4-E4B-it, the instruction-tuned E4B member of the Gemma 4 family. This is not a chat dataset or a benchmark. It is an internal-mechanics map built by running activations through a corpus of prompts and scoring what each layer, component, head, and feature direction is doing.
If you want to know what sliding-window and full-attention layers actually do differently inside one model, how KV cache sharing splits a… See the full description on the dataset page: https://huggingface.co/datasets/juiceb0xc0de/gemma-4-e4b-it-atlas.PraCegoVer-by-gemini3flash-corruptedcs2-v3-prompt-comparison-7-examples-with-multiaction-gemini35
CS2 V3 七案例 Gemini 3.5 Flash 最终结果对比 / Seven-case Gemini 3.5 Flash Comparison
本 README 展示 Gemini 3.5 Flash 对同一批 7 个视频的最终打标结果：每个案例先显示视频，再用左右两列并排展示两轮和七轮的完整 English JSON 与中文 JSON；内容直接展开，字号保持较小以便对照。
This README shows Gemini 3.5 Flash final labels for the same 7 videos. Each case places the video first, then displays complete English and Chinese JSON side by side: two-round on the left and seven-round on the right.
两轮与七轮的 API 输入详情通过顶部索引查看；中文侧保持与英文 JSON 相同的键、时间边界、数组长度和 Action 标签。
API… See the full description on the dataset page: https://huggingface.co/datasets/mikusama99/cs2-v3-prompt-comparison-7-examples-with-multiaction-gemini35.GameplayCaptions-Gemini-pro-visionSteamGlitches-Gemini-Labelspersian-ocr-gemini37-wins-bina-misses-viewer
Gemini 3.7 exact / Bina miss OCR crops
53 non-empty bbox crops from the PersianVLM submitted-10 benchmark where
google/gemini-3.7-flash was normalized-exact and Bina Koochik 0.1 was not.
This is a minimal Hugging Face ImageFolder dataset for reliable viewer support.
It contains exactly two columns: image and ocr. The ocr value is Gemini's
actual output for the corresponding crop.
persian-ocr-gemini37-wins-bina-misses-v2
Gemini 3.7 exact / Bina miss OCR crops
53 non-empty bbox crops from the PersianVLM submitted-10 benchmark where
google/gemini-3.7-flash produced a normalized-exact transcription and
Bina Koochik 0.1 did not.
This release uses the standard Hugging Face ImageFolder layout. The viewer's
first column is image, followed by ocr containing Gemini's actual output.
Gold and Bina outputs are retained for comparison.
gemma-4-e4b-it-500-refmhlc-training-gemma4-gemma4_e4b_it_think_on_hard_mixed_sources_120k
Multi Head Latent Control Training Data - Gemma 4 E4B it think on hard Mixed Sources 120k
Dataset Description
This repository contains verified training data for the Multi Head Latent Control paper release. It is part of the Multi Head Latent Control training data Hugging Face collection.
Paper
https://arxiv.org/abs/2607.14277
Code
https://github.com/Amirhosein-gh98/Multi-Head-Latent-Control
Dataset Summary
Field… See the full description on the dataset page: https://huggingface.co/datasets/AmirhoseinGH/mhlc-training-gemma4-gemma4_e4b_it_think_on_hard_mixed_sources_120k.maze-solving-for-gemma-4gemini-picture
Gemini.Picture 数据集
数据集位置
HuggingFace: https://huggingface.co/datasets/zhangtycx330/gemini-picture
本地路径: D:\Gemini.Picture\
数据集概览
2999 张图片，来自 4 个 HuggingFace 数据集
9 个语义类别（animal, architecture, general_mixed, landscape, portrait, scientific 等）
使用方法
从 HuggingFace 下载
from datasets import load_dataset
ds = load_dataset("zhangtycx330/gemini-picture")
使用本地数据
如果已下载到本地，修改 config/research.yaml 中的路径：… See the full description on the dataset page: https://huggingface.co/datasets/zhangtycx330/gemini-picture.gemini-3.7-flash-ocr-26-aug-2026
gemini-3.7-flash-ocr-26-aug-2026
Page-image → transcription pairs for finetuning a vision-language model to OCR
Devanagari and Tamil printed books.
These labels are not human ground truth. They are the output of a teacher
model, so its accuracy is the ceiling for anything trained on them.
Provenance
Teacher model
google/gemini-3.7-flash (via OpenRouter, reasoning.effort=low)
Page render
PyMuPDF at 200 DPI, grayscale JPEG q90
Sampling
stratified —… See the full description on the dataset page: https://huggingface.co/datasets/kailasa-ngpt/gemini-3.7-flash-ocr-26-aug-2026.Wait-Phenomenon-Evidence-Gemini-DeepSeekOriginal Repository: https://huggingface.co/datasets/hejun0180-pixel/Wait-Phenomenon-Evidence-Gemini-DeepSeek
Ordinary Agent — Meta-Learning — Autonomous Agent
Meta-Learning = Meta-Training = Meta-Social Agent + Civilization Meta-Rules + Emergent Tools
WP-AHA: Emergence Tool in LLMs
Attributes
Cross-Platform：Gemini 1.5 Pro, DeepSeek-V3/R1, Grok, GPT-4o, Claude, Doubao, Qwen, Kimi, Yuanbao.
Reproducible: Full Dataset ( 100+ WP-AHA—Endogenous Transition—Raw… See the full description on the dataset page: https://huggingface.co/datasets/hejun0180-pixel/Wait-Phenomenon-Evidence-Gemini-DeepSeek.gemstones_datasetmhlc-training-gemma4-gemma4_e4b_it_think_off_hard_mixed_sources_120k
Multi Head Latent Control Training Data - Gemma 4 E4B it think off hard Mixed Sources 120k
Dataset Description
This repository contains verified training data for the Multi Head Latent Control paper release. It is part of the Multi Head Latent Control training data Hugging Face collection.
Paper
https://arxiv.org/abs/2607.14277
Code
https://github.com/Amirhosein-gh98/Multi-Head-Latent-Control
Dataset Summary
Field… See the full description on the dataset page: https://huggingface.co/datasets/AmirhoseinGH/mhlc-training-gemma4-gemma4_e4b_it_think_off_hard_mixed_sources_120k.touch-ex
Touch-Ex: A Region-Level, Force-Annotated Visuo-Tactile Dataset
Developed by Gemma McLean and supervised by Dr Daniel Hao, University of Leicester.
Dataset Summary
Touch-Ex (Touch Exploration) is a region-level, force-annotated visuo-tactile dataset collected using a DIGIT vision-based tactile sensor and a collection of common UK household objects. The dataset was designed to support research in tactile representation learning, force-aware perception… See the full description on the dataset page: https://huggingface.co/datasets/gemixin/touch-ex.persian-ocr-gemini37-wins-bina-misses
Gemini 3.7 vs Bina OCR comparison crops
53 non-empty bbox crops where Gemini 3.7 Flash was normalized-exact and Bina
Koochik 0.1 was not. The first columns are image, gemini_ocr, bina_ocr,
and gold_text for direct visual and OCR comparison.
GEM-250K
GEM: Generative Supervision Helps Embodied Intelligence
Ruowen Zhao1,
Bangguo Li1,
Zuyan Liu1,2,†,
Yinan Liang1,
Junliang Ye1,
Fangfu Liu1,
Diankun Wu1,
Zhengyi Wang1,
Xumin Yu2,
Yongming Rao2,✉,
Han Hu2,
Jun Zhu1,✉
†Project Lead.✉Corresponding Author.
1Tsinghua University,
2Tencent Hunyuan
Abstract
Embodied Vision-Language Models (VLMs) have demonstrated impressive… See the full description on the dataset page: https://huggingface.co/datasets/zzzrw/GEM-250K.2026_07_19_collect_leandojo_gemma3_12b_gemma4_31bFinch-Collection-Gemini-3-Flash
Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks
A mid-training "practice phase" that teaches small open-source LLMs how to evolve solutions.
👋 This is the Gemini-3-Flash teacher variant of the Finch Collection — evolutionary search trajectories from the paper Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks, but with Gemini-3-Flash as the teacher mutation… See the full description on the dataset page: https://huggingface.co/datasets/minnesotanlp/Finch-Collection-Gemini-3-Flash.gemini-finetune-imagesPraCegoVer-by-gemini3flash