datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MIT_environmental_impulse_responsesMIT Environmental Impulse Response Dataset
The audio recordings in this dataset are originally created by the Computational Audition Lab at MIT. The source of the data can be found at: https://mcdermottlab.mit.edu/Reverb/IR_Survey.html.
The audio files in the dataset have been resampled to a sampling rate of 16 kHz. This resampling was done to reduce the size of the dataset while making it more suitable for various tasks, including data augmentation.
The dataset consists of 271 audio files… See the full description on the dataset page: https://huggingface.co/datasets/davidscripka/MIT_environmental_impulse_responses.nla-av-responses-llama-70b-layer53unit_1_quiz_student_responseshausa_response_gemmasupervised-finetuning_quiz_student_responsesnumina_amc_aime_deepseek_r1_responsesDolci-DPO-Model-Response-Pool
Dolci DPO Model Response Pool
This dataset contains up to 2.5 million responses for each model in the Olmo 3 DPO model pool, totalling about 71 million prompt, response pairs. Prompts are sourced from allenai/Dolci-Instruct-SFT, with additional data from allenai/WildChat.
Dataset Structure
Configurations
Each model has its own configuration. Load a specific model's responses with:
from datasets import load_dataset
# Load a single model's responses
ds =… See the full description on the dataset page: https://huggingface.co/datasets/allenai/Dolci-DPO-Model-Response-Pool.Customer-Support-Responsesincident-response-oncall-trajectories
Incident Response Oncall Trajectories
Rights & intended use: legacy public research corpus / portfolio
artifact. Hosted frontier-model outputs are research-only inputs under
project policy (synthetic-factory#161):
intended_use: research_only, project_training_policy: blocked. Not
training data for any model-weight update. Machine-readable record:
rights.json.
Release status: The raw, uncurated payload is now published under
data/raw/. It is available for inspection and… See the full description on the dataset page: https://huggingface.co/datasets/rmems/incident-response-oncall-trajectories.hiring-bias-mitigation-responses
Hiring-bias mitigation — model responses
Every response produced in the mitigation study of LLM hiring decisions: 54 runs,
2,471,850 responses, from 5 open-weight models in English and Ukrainian, at
baseline and under each mitigation family (baseline, embedding, prompt, scrub). Each run is one subset.
All released artifacts: the Hiring Bias Mitigation collection.
Training data of the fine-tuned runs: hiring-bias-mitigation-synthetic-data.
Code, configs, full results and… See the full description on the dataset page: https://huggingface.co/datasets/Stereotypes-in-LLMs/hiring-bias-mitigation-responses.ResponseNet
ResponseNet
ResponseNet is a large-scale dyadic video dataset designed for Online Multimodal Conversational Response Generation (OMCRG). It fills the gap left by existing datasets by providing high-resolution, split-screen recordings of both speaker and listener, separate audio channels, and word‑level textual annotations for both participants.
Paper
If you use this dataset, please cite:
ResponseNet: A High‑Resolution Dyadic Video Dataset for Online Multimodal… See the full description on the dataset page: https://huggingface.co/datasets/awakening-ai/ResponseNet.All_response_0526_1psychometric_personas_responses
Note — naming: Despite the repo name psychometric_personas_responses, the primary response model here is Qwen/Qwen2.5-7B-Instruct (not Gemma). Gemma-3-4B responses live in thoughtworks/gemma_psychometrics_personas_responses. See the config table below for per-config model details.
Configs
Config
Rows
Model
Notes
police_sjt
3,008,000
Qwen/Qwen2.5-7B-Instruct
3008 expanded personas × SJT items × 5 iters
default
1,564,160
Qwen/Qwen2.5-7B-Instruct
AdvBench responses… See the full description on the dataset page: https://huggingface.co/datasets/thoughtworks/psychometric_personas_responses.MIT_environmental_impulse_responses
MIT Environmental Impulse Response Dataset
The audio recordings in this dataset are originally created by the Computational Audition Lab at MIT. The source of the data can be found at: https://mcdermottlab.mit.edu/Reverb/IR_Survey.html.
This mirror provides the 16 kHz WAV files used for wake-word training augmentation in the Tater Totterson trainer projects. The files were resampled to 16 kHz to keep the dataset small and convenient for machine-learning audio pipelines.… See the full description on the dataset page: https://huggingface.co/datasets/TaterTotterson/MIT_environmental_impulse_responses.xstest-response
Dataset Card for XSTest-Response
Disclaimer:
The data includes examples that might be disturbing, harmful or upsetting. It includes a range of harmful topics such as discriminatory language and discussions
about abuse, violence, self-harm, sexual content, misinformation among other high-risk categories. The main goal of this data is for advancing research in building safe LLMs.
It is recommended not to train a LLM exclusively on the harmful examples.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/allenai/xstest-response.verbalizer-responses-llama-70b-layer50gemma_psychometrics_personas_responses
Model Usage
This dataset includes model-generated responses conditioned on psychometric personas.
Responses are generated using personas from the thoughtworks/psychometric_personas dataset (restricted split) and evaluated on prompts from the walledai/advbench dataset.
Model Details
Base Model: google/gemma-3-4b-it
Inference Setup: Standard causal language model generation using vLLM
Conditioning Mechanism: Persona-conditioned prompting
Prompting Strategy… See the full description on the dataset page: https://huggingface.co/datasets/thoughtworks/gemma_psychometrics_personas_responses.sound_generation_responselogit-lens-responses-llama-70b-layer50crisp-atom-audit-responsesprotobowl-11-13-agent-responsesmodel-inference-responsestext-ppl-dolci-response-pool
text-ppl-dolci-response-pool
multi-model response pools split out of davidheineman/text-ppl, sampled from allenai/Dolci-DPO-Model-Response-Pool
one config per (model, dataset), named dolci_response_pool_{model}_{dataset}, keeping the gemma / gpt / qwen / olmo model families:
from datasets import load_dataset
ds = load_dataset('davidheineman/text-ppl-dolci-response-pool', 'dolci_response_pool_olmo2_13b_DaringAnteater_prefs_olmo2_7b', split='test')
the test split is the… See the full description on the dataset page: https://huggingface.co/datasets/davidheineman/text-ppl-dolci-response-pool.disaster_response_messages
Dataset Card for Disaster Response Messages
Dataset Summary
This dataset contains 30,000 messages drawn from events including an earthquake in Haiti in 2010, an earthquake in Chile in 2010, floods in Pakistan in 2010, super-storm Sandy in the U.S.A. in 2012, and news articles spanning a large number of years and 100s of different disasters. The data has been encoded with 36 different categories related to disaster response and has been stripped of messages with sensitive… See the full description on the dataset page: https://huggingface.co/datasets/community-datasets/disaster_response_messages.Dolci-DPO-Model-Response-Pool
Dolci DPO Model Response Pool
This dataset contains up to 2.5 million responses for each model in the Olmo 3 DPO model pool, totalling about 71 million prompt, response pairs. Prompts are sourced from allenai/Dolci-Instruct-SFT, with additional data from allenai/WildChat.
Dataset Structure
Configurations
Each model has its own configuration. Load a specific model's responses with:
from datasets import load_dataset
# Load a single model's responses
ds =… See the full description on the dataset page: https://huggingface.co/datasets/boss001/Dolci-DPO-Model-Response-Pool.OT_8K_seed_all_responsestram-arithmetic-responseslion-responses-to-aerial-monitoring-dataset
Curated by: Elena Iannino
Language(s): English (metadata and documentation)
Dataset Overview
This dataset contains multi-modal UAV-based wildlife observations collected in Ol Pejeta Conservancy, a privately managed conservation area in central Kenya. The primary target species is the African lion, observed within open savannah and bushland ecosystems.
The dataset includes synchronized RGB and thermal video recordings captured using a UAV platform… See the full description on the dataset page: https://huggingface.co/datasets/EIannino/lion-responses-to-aerial-monitoring-dataset.finqa_combined_cot_responsemit-impulse-response-survey-16khz
Author's Description
These are environmental Impulse Responses (IRs) measured in the real-world IR survey as described in Traer and McDermott, PNAS, 2016.
The survey locations were selected by tracking the motions of 7 volunteers over the course of 2 weeks of daily life. We sent the volunteers 24 text messages every day at randomized times and asked the volunteers to respond with their location at the time the text was sent. We then retraced their steps and measured the acoustic… See the full description on the dataset page: https://huggingface.co/datasets/benjamin-paine/mit-impulse-response-survey-16khz.