datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
friend-bench
Can a model — or a human — tell how two people are related from a 20-second clip of how they interact?
🌐 Built on Seamless Interaction
FriendBench is a suite of benchmarks for social perception from thin-slice dyadic
interaction — inferring facts about two people's relationship from a brief clip of how they
interact, built on the Seamless Interaction
dataset. Each released set is a config of this repository.
🎧 Multi-modal — text, audio, and video for every clip
🎯 Objective label —… See the full description on the dataset page: https://huggingface.co/datasets/fluid-concepts/friend-bench.human-conceptsDataset Summary
This dataset provides digitized versions of classic human categorization benchmarks from seminal cognitive psychology studies by Rosch (1973, 1975) and McCloskey & Glucksberg (1978). These datasets capture human judgments about semantic categories and typicality, offering high-fidelity insights into how humans organize conceptual knowledge.
This dataset was released as part of the study "From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning"
(Shani et al.… See the full description on the dataset page: https://huggingface.co/datasets/CShani/human-concepts.openalex-concepts
OpenAlex L1 + L2 Concepts
What this is
A snapshot of OpenAlex's Level 1 (broad fields) and Level 2 (subfields) concepts: 21,739 records across 284 broad fields and 21,455 subfields.
Level
Count
Examples
1
284
Computer science, Physics, Biology, Sociology
2
21,455
Machine learning, Quantum mechanics, Convolutional neural network
Each record:
Column
Type
Description
id
string
OpenAlex concept ID (e.g., C121955636)
name
string
Concept… See the full description on the dataset page: https://huggingface.co/datasets/barissozudogru/openalex-concepts.mining_conceptsconcept-sentimentmining_concepts_special_format_trainingmining_concepts_format_validation