Base_Ft
Datasets
All datasets matching “Base_Ft”maxrl_qwen3_4B_base_polaris_rollouts
MaxRL Qwen3-4B-Base training rollouts (POLARIS math prompts)
Every training rollout from an online RL run, with exact token ids, sampling
log-probs, and raw rewards — usable as a replay buffer to study off-policy RL
for LLM reasoning completely offline.
The run: Qwen3-4B-Base trained with the maxRL advantage estimator
(A = (r - mean)/(mean + eps), group mean over 16 rollouts per prompt;
maxRL paper) and a pure REINFORCE loss
(L = -A * log pi; no importance ratio, no clipping, no… See the full description on the dataset page: https://huggingface.co/datasets/ftajwar/maxrl_qwen3_4B_base_polaris_rollouts.cond_ft_none_on_reddit__prcnt_100__test_run_False__xlm-roberta-basecond_ft_subreddit_on_reddit__prcnt_100__test_run_False__xlm-roberta-baseOlmoEarth-v1-FT-Karst-Groundwater-Base
OlmoEarth-v1-FT-Karst-Groundwater-Base
Fine-tuning dataset for OlmoEarth-v1-Base to detect karst terrain features in Pennsylvania from Sentinel-1 SAR backscatter and Sentinel-2 optical imagery.
47,020 labeled points · 4 classes · ~50,000 km² · 3 AOI regions
Quick Start
Two formats are provided for the two upload paths:
Option A — OlmoEarth Studio Web UI (studio/import.geojson)
Upload via Studio's "Import your own dataset" feature. Map the properties to your… See the full description on the dataset page: https://huggingface.co/datasets/2imi9/OlmoEarth-v1-FT-Karst-Groundwater-Base.cond_ft_none_on_reddit__prcnt_na__test_run_True__bert-base-uncasedcond_ft_none_on_reddit__prcnt_100__test_run_False__bert-base-uncased