Sign in to view source links and access this dataset
Description
PaperFlow-Bench is a dataset of simulated user interactions with academic papers, packaged as a Hugging Face dataset by OpenRaiser. It includes user metadata, daily interaction episodes, paper metadata from arXiv, and labels for reading selections. The dataset was last updated on June 4, 2026.
Use Cases
Training recommender systems based on simulated user profiles and interaction episodes.
Evaluating recommendation algorithms using the provided episode-paper labels and reading selections.
Studying temporal user interest drift patterns based on the diagnostic timeline data.
Analyzing paper metadata and abstracts for content-based filtering models.
Strengths
Contains multiple linked data files (users, episodes, papers, labels, timeline) for a structured simulation.
Paper metadata includes arXiv abstracts and PDF URLs, providing access to source content.
Includes a specific diagnostic timeline for analyzing interest drift.
Limitations
Row counts and dataset size are unknown, which may limit suitability assessment.
Column-level documentation is absent; field semantics must be inferred after download.
Data is simulated, which may not fully reflect real-world user behavior patterns.
Provenance
Source
OpenRaiser via Hugging Face.
Collection Method
Simulated user metadata and interactions; paper metadata sourced from arXiv.
Freshness
Last updated 2026-06-04 08:52:29.
License information is unknown; users should verify terms of use before application.