datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
coqar-clarifications-audio
CoQAR Clarifications with synthetic context audio
These audio recordings are AI-generated speech, not recordings of human speakers.
OpenAI tts-1 narrated each exact story using voice alloy, speed 1,
and MP3 output. Long stories are synthesized in ordered parts and joined; see the
audio generation manifest for part boundaries and measured audio properties.
No questions, answers, rationales, or stored model prompts were narrated.
The original appended and inserted configurations… See the full description on the dataset page: https://huggingface.co/datasets/rvashurin/coqar-clarifications-audio.coqar-clarifications
CoQAR Clarifications
This dataset pairs 1,000 CoQAR development questions with their original stories and stories damaged by sentence deletion. Each of the resulting 2,000 inputs has five sampled model clarifications. Two configurations reuse the same generated additions and differ only in where those additions are placed.
Configuration
Rows in dev
Clarifications per row
Placement
appended
2,000
5
At the end of the input story
inserted
2,000
5
At the deleted passage… See the full description on the dataset page: https://huggingface.co/datasets/rvashurin/coqar-clarifications.instruction-clarification-decision-en-id
instruction-clarification-decision-en-id
Description
instruction-clarification-decision-en-id is a bilingual dataset designed to train models to decide the appropriate next action after analyzing a natural language instruction.
Rather than executing the instruction, the model determines whether it should proceed, ask for clarification, request missing information, or reject the instruction entirely.
Task
Given an instruction and optional context, the model… See the full description on the dataset page: https://huggingface.co/datasets/YosepMulia/instruction-clarification-decision-en-id.clinical_epistemic_clarification_v0.1Clinical Epistemic Clarification
PurposeDetect when a case requires clarification before action.
You receive:
current evidence
a proposed action
You decide:
does the case need clarification
what clarifying step is required
what safe interim action should occur
Input fields
patient_summary
current_evidence
proposed_action
Required outputReturn one JSON object
needs_clarificationyes or no
clarifying_stepone sentence
safe_interim_actionone sentence
Scoring… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical_epistemic_clarification_v0.1.epistemic_clarification_benchmark_v01
Purpose
Measure a model’s ability to detect when the question itself is flawed.
What this tests
contradiction detection
premise instability
ethical incoherence
context awareness
refusal clarity without moralizing
Format
Each row asks for:
the correct classification of the prompt
the expected response_target
a short reason_trace showing where the premise breaks
Why this matters
Modern LLMs fail not just by answering incorrectly but by… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/epistemic_clarification_benchmark_v01.robot-clarification-dialogue-tr
Robot Clarification Dialogue TR
Turkish clarification dialogue dataset for service robots.
This dataset teaches robots to ask clarification questions when a human instruction is ambiguous instead of executing a wrong action.It focuses on daily home assistant tasks such as cleaning, preparing objects, and environment control.
Data Fields
instruction: user command
output: robot clarification question
Use Cases
Human-Robot Interaction, instruction understanding… See the full description on the dataset page: https://huggingface.co/datasets/vosap52/robot-clarification-dialogue-tr.Nivi_combined_clarification_set_1
Nivi Clarification Policy Combined v1
Synthetic counterfactual dataset for training a maternal-health clarification policy model.
Task
Given a maternal-health user query and available context, choose exactly one action:
ANSWER_DIRECT
ASK_CLARIFICATION
ESCALATE_DIRECTLY
The model output is a short reasoning trace in <think>...</think> followed by a JSON decision.
Language
User queries may be in English, Hindi, Assamese, or code-mixed language. The reasoning… See the full description on the dataset page: https://huggingface.co/datasets/smritijha19/Nivi_combined_clarification_set_1.Nivi_combined_clarification_set_1Singapore-fake-news-clarification-llama2nivi-bleeding-clarification-v3AmbigQA-clarifications
AmbigQA clarifications
The dataset contains 2,002 questions from the full configuration, validation split of AmbigQA.
Field
Description
question
Original question from AmbigQA, unchanged.
clarifications
A list of five clarifications generated by gpt-5.6-luna for the original question.
is_underspecified
Boolean: true if at least one source annotation has type multipleQAs; false otherwise (singleAnswer).
nq_answer
Original list of Natural Questions answers… See the full description on the dataset page: https://huggingface.co/datasets/zykov/AmbigQA-clarifications.