datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
4b_rft_response-2-custom_student_response4b_rft_response-7-custom_student_response-verified4b_rft_response-5-custom_student_response-verified4b_rft_response-3-custom_student_response-verified4b_rft_response-4-custom_student_response-verified4b_rft_response-1-custom_student_response-verifiedlibero-rft-spatial-full14b_rft_response-4-custom_student_response4b_rft_response-2-custom_student_response-verifiedkk-onpolicy-rftkk-rft-trainlibero-rft-spatial-459-wide4b_rft_response-6-custom_student_response-verified-accrf_train_200_1SATQuest-RFT-3k
SATQuest Dataset
TL;DR. Synthetic CNF benchmark for LLM reasoning: 3000 matched SAT/UNSAT pairs with n in [3, 8] and ratio m in [2.1n, 4n]. The dataset stores only CNF formulas and solver stats; use the SATQuest Python library to render prompts/answers for SATDP, SATSP, MaxSAT, MCS, and MUS in four formats (math, DIMACS, story, dual story).
Data fields
id: unique identifier for each row.
num_literal: total number of literals in the unsatisfiable formula.
sat_dimacs:… See the full description on the dataset page: https://huggingface.co/datasets/sdpkjc/SATQuest-RFT-3k.4b_rft_response-5-custom_student_response-verified-acc4b_rft_response-7-custom_student_response-verified-acccra-ccf_RFT4b_rft_response-2-custom_student_response-verified-acc4b_rft_response-3-custom_student_response-verified-acc4b_rft_response-1-custom_student_response-verified-accqwen3-4b-hard-math-mix-guided-full-rftMM_RFTautoteacher-rft-base4b_rft_response-4-custom_student_response-verified-accvimedaqa-rft-poolautoteacher-rft-base-allSATQuest-RFT-1k4b_rft_response_acc_rolloutY_mixautoteacher-rft-hard1k