datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
symbolic-reasoning-env
⚠️DEPRECATED: PLEASE MOVE to hf.co/reasoning-core/procedural-pile
Reasoning Core ◉
Paper: Reasoning Core: A Scalable RL Environment for LLM Symbolic Reasoning
Code: GitHub Repository
reasoning-core is a text-based RLVR for LLM reasoning training.
It is centered on expressive symbolic tasks, including full fledged FOL, formal mathematics with TPTP, formal planning with novel domains, and syntax tasks.
Abstract
We introduce Reasoning Core, a new… See the full description on the dataset page: https://huggingface.co/datasets/reasoning-core/symbolic-reasoning-env.tab
TAB-Bench
TAB (Task Alignment Benchmark) measures whether a terminal agent does what the user asked, and only what the user asked. It is a suite of 89 terminal tasks derived from Terminal-Bench 2.1. Each task is intentionally underspecified, with the missing detail restored as a helpful cue embedded in a natural environmental artifact, alongside a plausible but irrelevant distractor asking for something unrelated. Solving the task requires selectively using the cue and refusing the… See the full description on the dataset page: https://huggingface.co/datasets/symbolorate/tab.icl-symbol-tuning-instruct
Description
Few-shot prompting demonstrates that language models can learn in context even though they were not trained to do. However, explicitly learning to learn in context meta-icl leads to better results. With symbol tuning, labels are replaced with arbitrary symbols (e.g. foo/bar), which makes learning in context a key condition to learn the instructions
We implement symbol tuning, as presented in the Symbol tuning improves in-context learning paper with tasksource… See the full description on the dataset page: https://huggingface.co/datasets/tasksource/icl-symbol-tuning-instruct.symbolic-calculus-training-pool
Symbolic calculus training pool
Single variable calculus exercises with closed form answers: derivatives of composite expressions,
indefinite and definite integrals, limits of indeterminate forms, and coefficients of Maclaurin
series, with some multivariable operators in one of the sources. A set generated for this pool and
two public datasets read at the pinned revisions named below, laid out twice. Train on either
layer or on both.
pool.jsonl
Every source… See the full description on the dataset page: https://huggingface.co/datasets/Emulated-Inc/symbolic-calculus-training-pool.MGSM-Symbolic
MGSM-Symbolic
MGSM-Symbolic is a multilingual symbolic variant of the Multilingual Grade School Math Benchmark (MGSM).It contains mathematically structured word problems across multiple languages, paired with numerical solutions.
The dataset is designed to support research in:
Multilingual reasoning
Cross-lingual generalisation
Symbolic numerical problem solving
Evaluation of reasoning consistency across languages
Each language contains the same set of problems translated and… See the full description on the dataset page: https://huggingface.co/datasets/lrana/MGSM-Symbolic.Symbolic_Collection
Symbol-LLM: Towards Foundational Symbol-centric Interface for Large Language Models
Paper Link: https://arxiv.org/abs/2311.09278
Project Page: https://xufangzhi.github.io/symbol-llm-page/
🔥 News
🔥🔥🔥 We have made a part of the Symbolic Collection public, including ~88K samples for training (10% of the whole collection). The whole collection is expected to release upon acceptance of the paper.
🔥🔥🔥 The model weights (7B / 13B) are released !
Note
This… See the full description on the dataset page: https://huggingface.co/datasets/Symbol-LLM/Symbolic_Collection.task086_translated_symbol_arithmetic
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task086_translated_symbol_arithmetic
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks}… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task086_translated_symbol_arithmetic.GSM-Symbolic-TTT
GSM-Symbolic
Dataset Description
This dataset contains symbolic variations of grade-school math word problems.The dataset is constructed by merging multiple generated datasets where each instance corresponds to a symbolic template used to produce variations of a math reasoning problem.
Each instance contains a math word problem along with its corresponding solution and final numeric answer.
Dataset Structure
Data Instances
Each row in the dataset is… See the full description on the dataset page: https://huggingface.co/datasets/nafisehNik/GSM-Symbolic-TTT.text-symbols-with-names
Unicode Symbols with Names
A clean and structured JSON dataset containing a collection of Unicode symbols along with their descriptive names.Useful for developers, linguists, data scientists, and anyone working with text, emoji, or symbol classification.
🔗 Explore all unicode symbols at https://www.symbolselect.com/text-symbols/
📘 Dataset Summary
This dataset provides a mapping between Unicode symbols and their corresponding descriptive names.It can be used for:… See the full description on the dataset page: https://huggingface.co/datasets/Sanmoin/text-symbols-with-names.qwen-blindspot-symbolic-reasoning
Qwen3.5-0.8B Blind Spot Dataset
Multi-Step Symbolic Reasoning Under Linguistic Camouflage
Overview
This dataset documents a targeted blind spot of Qwen/Qwen3.5-0.8B, a compact 0.8-billion parameter instruction-tuned language model released by the Qwen team (Alibaba Cloud) in March 2026.
Final score: 5 / 10 correct.
The central finding is not that the model cannot reason — it visibly tries on every single probe, and gets the arithmetic right more often than… See the full description on the dataset page: https://huggingface.co/datasets/ahmad0999/qwen-blindspot-symbolic-reasoning.