AI BriefingPaper Discovery Tool Tomesphere Launches
Key point
Tomesphere, a research tool offering TLDRs, reviews, code, and semantic similarity graphs for 3 million papers, has launched.
Details
Tomesphere, a web service providing rich metadata for 3 million papers from arXiv and OpenAlex, has been released. It is available for free without any sign-up or payment required.
Each paper page integrates the following key information:
- A Gemini-based summary (TLDR)
- Peer reviews and decisions collected from OpenReview
- Related GitHub repositories, HuggingFace models, and datasets
- Conference videos and an OpenAlex-based citation graph
Notably, unlike existing tools such as Connected Papers or Litmaps that rely on citation overlap, Tomesphere defaults to text vector similarity using SPECTER2. This makes it effective at finding recent papers that have not yet formed citation relationships, or research that is topically similar despite having a different citation network.
Users can explore papers through four ranking modes—Influential, Recent, Hidden gems, and Nearest—and can view the data directly on arXiv pages via a Chrome extension.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Scaling Domain Data Repetition in LLM Pre-trainingResearchers Release Git-Based AI Research Agent Platform 'Agora'… Closes 62% of GPT-2 Gap Without Training