datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
OracleProto
OracleProto: Forecasting Evaluation Set
Chinese doc: [中文文档]
GitHub repo: [MaYiding/OracleProto]
Visit Our Leaderboards: [Website]
View Our Paper: [arXiv]
A SQLite-packaged evaluation set of 80 hand-curated forecasting questions on real-world events, with resolution dates between 2026-03-12 and 2026-04-14, released alongside the GitHub Repo. Both the rows and the byte-stable prompt-reconstruction recipe are packaged in a single file, forecast_eval_set_example.db, which exposes two… See the full description on the dataset page: https://huggingface.co/datasets/MaYiding/OracleProto.if_oracle_sft
IF Oracle SFT
Oracle instruction-following SFT data built by rejection sampling from
willhx/Qwen3-8B-Base-IF.
How it was made
Policy: willhx/Qwen3-8B-Base-IF (a trained IF RL checkpoint).
Prompts: the full allenai/IF_multi_constraints_upto5-derived training set
(IF_multi_constraints_upto5_ifbench_en, 88,832 prompts covered).
Sampling: 8 responses per prompt, temperature 1.0, max 8192 response tokens
(rollout via slime + SGLang, --debug-rollout-only).
Reward:… See the full description on the dataset page: https://huggingface.co/datasets/willhx/if_oracle_sft.cot-oracle-data
CoT Oracle Training Data
Datasets for training an Activation Oracle to detect unfaithful chain-of-thought reasoning.
Splits
thought_anchors (3 examples)
Qwen3-8B CoT traces with resampled importance scores per sentence. Each problem was solved 5 times with the CoT truncated at different points to measure which sentences causally influence the answer.
Fields:
sentences[i].importance_delta: how much removing this sentence hurts accuracy (0-1)… See the full description on the dataset page: https://huggingface.co/datasets/ceselder/cot-oracle-data.text2sql-oracle-postgres
Oracle / PostgreSQL text-to-SQL
Instruction data for fine-tuning google/gemma-3-270m-it (or any chat model) to emit a single dialect-correct SQL statement.
804 rows, 402 Oracle / 402 PostgreSQL
7 schemas: hr, sales, banking, inventory, tickets, university, logistics
Splits: 684 / 60 / 60 (grouped so paraphrases of the same SQL stay in one split)
Load
from datasets import load_dataset
ds = load_dataset("chabab/text2sql-oracle-postgres")
Record… See the full description on the dataset page: https://huggingface.co/datasets/chabab/text2sql-oracle-postgres.experiment-1-oracles
Experiment 1 Oracles Corpus
Dataset Summary
Dataset containing synthetic ground-truth oracle responses and intermediate reasoning traces for multi-step AI experiment evaluation.
personahub-lora-oracles
PersonaHub LoRA Oracles
25,000 rank-1 LoRA weight diffs trained on PersonaHub personas on Gemma 3 4B-it. Designed for training weight oracle models that can read LoRA weights and describe model behavior.
Dataset Structure
personas.parquet — 25K persona descriptions (persona_id, persona_text)
dataset_config.json — Full reproduction config (model, hyperparams, prompts)
loras/ — Raw LoRA weight files (.pt), one per persona (~3.75MB each, fp16)
direction_tokens/ —… See the full description on the dataset page: https://huggingface.co/datasets/ceselder/personahub-lora-oracles.