Twitxy Saved

Benjamin Feuer

@FeuerBenjamin

PhD Candidate in Computer Science, NYU, Deep Learning
New York Metro Joined May 2022
154 Following 139 Followers 180 Posts

Latest posts 5

Benjamin Feuer @FeuerBenjamin
So excited to be the SFT lead of this massive collaboration! The OpenThoughts team is πŸ”₯
Negin Raoof Verified account @NeginRaoof_ Dec 6, 2025
How can we make a better TerminalBench agent?
Today, we are announcing the OpenThoughts-Agent project.
OpenThoughts-Agent v1 is the first TerminalBench agent trained on fully open curated SFT and RL environments.
OpenThinker-Agent-v1 is the strongest model of its size on TerminalBench, and sets a new bar on our newly released OpenThoughts-TB-Dev benchmark. (1/n)
View quoted post on X
3 replies 1 reposts 14 likes 1.5K views View on X
Benjamin Feuer @FeuerBenjamin
New research paper for you to read over your July 4th break (if you're US-based) --

Vision is a skeleton key! πŸ—οΈ We convert a small VLM into an "everything classifier" by transforming data into visualizations that VLMs can naturally understand and reason about. We call it MARVIS: Modality Adaptive Reasoning over VISualizations.

Our MARVIS-3B model:
- Beats Gemini by 16% on average across 100s of vision and tabular tasks πŸ†
- Gets within 2.5% of the best specialized model across across 4 modalities ... 🎯
- Using just one 3B model ... πŸ’ͺ
- ... without exposing any P.I.I. (personally identifiable information) to the VLM ... πŸ”
- And without requiring any model training! ⚑

Our GitHub: lnkd.in/eqPkzU2m πŸ’»
Our Paper: lnkd.in/esXZEvEE πŸ“„
Research Supported By: oumi.ai
Thanks to @LennartPurucker @Oussama_e
Benjamin Feuer @FeuerBenjamin
So excited to announce the DCVLR (Data Curation for Vision-Language Reasoning) competition at NeurIPS 2025, led by @Oumi_PBC and sponsored by @LambdaAPI!

🌟open-data 🌟
πŸ€– open-models πŸ€–
πŸ’» open-source πŸ’»
πŸ’ͺanyone can compete for free πŸ’ͺ

dcvlr-neurips.github.io

🧡 1 / n
dcvlr-neurips.github.io DCVLR: NeurIPS 2025 Competition Join the DCVLR NeurIPS 2025 Competition. Advance visual reasoning in VLMs through data curation.
1 replies 14 reposts 43 likes 11K views View on X
Benjamin Feuer @FeuerBenjamin
To all the AI people remembering fondly they were on a paper with Geoffrey Hinton, or a panel with Geoffrey Hinton, or accidentally sneezed into a handkerchief held by Geoffrey Hinton...

Geoffrey Hinton won a Nobel prize. You did not. Stop exploiting this for clicks.