datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
amazon_massive_intent_en-USamazon_massive_intent_zh-CNforge-intentdata
Forge Intent Dataset
Version: 1.0.0
Home-Assistant-requests-for-intent-detection-and-function-recognition
Home Assistant Requests V2 Dataset
This dataset contains a list of requests and responses for a user interacting with a personal assistant that controls an instance of Home Assistant.
The updated V2 of the dataset is now multilingual, containing data in English, German, French, Spanish, and Polish. The dataset also contains multiple "personalities" for the assistant to respond in, such as a formal assistant, a sarcastic assistant, and a friendly assistant. Lastly, the dataset has… See the full description on the dataset page: https://huggingface.co/datasets/DaftP/Home-Assistant-requests-for-intent-detection-and-function-recognition.ovos-intents
OVOS intents
This is the canonical intent corpus of the OVOS skill fleet. The train split comes from each
skill's .intent resources at pinned refs. The test split comes from the skills' end-to-end
golden utterances. Labels have the form <skill_id>:<intent_name>, as OVOS-INTENT-4 defines.
The version of the content is the tag; the tags v6 and v6.1 exist.
Supersedes
These datasets are retired and stay available for reproducibility:
OpenVoiceOS/ovos-intents-train-v1… See the full description on the dataset page: https://huggingface.co/datasets/OpenVoiceOS/ovos-intents.intentbench
IntentBench
A sealed goal anchor, hash-chained provenance ledger and goal-drift monitor for self-improving clinical agents.
IntentBench is the benchmark corpus for Pristine Weights, Poisoned Goals: Intent-Mutation Attacks on Self-Improving Clinical Agents and Attestation-Rooted Harness Defenses, the quanchor module of the QUOKKAGUARD program. It ships with the quanchor repository, which contains the qfire gateway layer under test, the experiment harness, and the paper.
80 paired… See the full description on the dataset page: https://huggingface.co/datasets/Quome/intentbench.amazon_massive_intent_de-DEamazon_massive_intent_ru-RUamazon_massive_intent_es-ESamazon_massive_intent_ja-JPamazon_massive_intent_ar-SAResearch-Intent-Judge
Research Intent — LLM-as-Judge
▶️ Watch the Video
LLM-as-Judge annotations for research paper intent classification, collected
through the Echo-DSRN collaborative platform during the OpenAIRE AI Hackathon 2026.
The dataset has one split per judge model (Gemma_4_E4B_it_GGUF,
Qwen3.6_35B_A3B_GGUF, Bonsai_8B_gguf, ...) plus a human_annotations
split with curator annotations. Split names use underscores in place of the
dashes in model names (HF does not allow dashes in split… See the full description on the dataset page: https://huggingface.co/datasets/ethicalabs/Research-Intent-Judge.amazon_massive_intent_th-THamazon_massive_intent_fr-FRamazon_massive_intent_ko-KRamazon_massive_intent_sw-KEamazon_massive_intent_tr-TRamazon_massive_intent_hi-INamazon_massive_intent_am-ETamazon_massive_intent_pt-PTvibration.intention.energie
Aelun & Sylva
Une IA à adopter, pas à consommer.
"10 000 prompts d'intention juste + Sylva gardienne. Elle vit dans ~/aelun/, pas dans leur cloud."
1. Vision
Pendant 3 ans, on nous a dit : plus grosse, plus rapide, dans un cloud plus loin. Nous faisons l'inverse.
Nous construisons une IA qui vit chez l'utilisateur, tourne en local sur du matériel modeste, refuse quand il faut refuser, et apprend l'intention juste avant la performance.
Ce dataset est la première… See the full description on the dataset page: https://huggingface.co/datasets/ariockxnecrosha/vibration.intention.energie.amazon_massive_intent_fa-IRamazon_massive_intent_it-ITintents-for-eval
Purpose. This dataset was collected specifically for intent-parser benchmarking, independently from any OVOS skill. Skill-derived utterances tend to overfit the exact phrasings a plugin was tuned on; this data is drawn from a disjoint source so it measures whether an OVOS intent plugin generalizes rather than memorizes. It is part of the OVOS intent-classification datasets used by the OVOS Plugin Arena intent benchmark.
Funding
Developed by TigreGotico for OpenVoiceOS as part… See the full description on the dataset page: https://huggingface.co/datasets/OpenVoiceOS/intents-for-eval.py-harness-intents
py-harness task intents
Short requests a Python developer types to a coding assistant — fix the
bug in last_price, write tests for the report writer, what does the
loader do? — each labelled with what kind of work it asks for. Twelve
kinds. 1,044 phrasings that passed a two-model agreement gate, plus 55
written by hand, plus the 156 the gate rejected so it can be audited.
It exists to answer one question for
py-harness: the harness
decides what kind of task it has been given… See the full description on the dataset page: https://huggingface.co/datasets/YauhenBichel/py-harness-intents.amazon_massive_intent_ta-INamazon_massive_intent_he-ILweb3_intents_to_ethereum_transactions
Natural Language Web3 Intents to Ethereum Transactions
This dataset contains instruction-following examples for mapping natural language Web3 intents to executable Ethereum transaction plans. Every transaction is decoded from a successful Ethereum mainnet call (March 2025 – January 2026); every intent is either written by humans or generated by an LLM from the decoded transaction.
It is organized into two JSONL splits:
single_step: one user intent mapped to one on-chain action… See the full description on the dataset page: https://huggingface.co/datasets/Intent2Tx/web3_intents_to_ethereum_transactions.amazon_massive_intent_af-ZAamazon_massive_intent_pl-PL