Hugging Face's logo
Reuben fernandes's picture
šŸ—ļø Building on HF

Reuben fernandes PRO

Reubencf
hugging-science
28 42 136
Akhil-Theerthala's profile picture SelmaNajih001's profile picture KennyVechov's profile picture
Ā·
http://reuben-fernandes.xyz/

AI & ML interests

LLM

Recent Activity

updated a Space about 13 hours ago
Reubencf/HF_QR_Gen
reacted to RDTvlokip's post with šŸ‘ 2 days ago
I finally changed the architecture of my 15M French LLM. It worked. Then I almost fooled myself about how much and catching that was the real win. After proving last time that architecture is a threshold, not a lever, I got stubborn: could I change how the model learns? Four honest attempts, Lion, a sharper AdamW Ī²2, multi-token prediction, LayerScale. Four failures. The bottleneck wasn't the learning rule either. So I changed the shape of the computation instead: loop the same transformer blocks 4Ɨ, deeper reasoning, zero added parameters. It beat the baseline on perplexity, the first thing in the whole project to move that number. Then I added my own twist: let each token decide how deep to think, halting on its own entropy. My first evaluation was spectacular. Coherence up 65%. Hallucinated names down 62%. It was noise. Eight prompts, one seed. I re-ran on 50 prompts Ɨ 200 tokens and watched the gains shrink to "modest" and on out-of-domain prompts, recurrence actually made things worse. No universal winner. And none of it is new: it's Adaptive Computation Time (2016), the Universal Transformer (2018), and LoopViT (2026), recombined and measured honestly. The real lesson: A number from 8 prompts is a rumor. The eval harness that kills your own best result is worth more than the result it kills. Cite your lineage. Stay preliminary until multiple seeds say otherwise. The three models are live. The write-up is honest about every caveat šŸ‘‡ šŸ”— https://huggingface.co/blog/RDTvlokip/teaching-a-15m-french-llm-to-think-deeper
reacted to theirpost with šŸ§  2 days ago
Introducing https://huggingface.co/spaces/Reubencf/Document_Query powered by https://huggingface.co/CohereLabs/command-a-plus-05-2026-w4a4 by cohere just open admin page -> drop in any document -> and ask questions
View all activity

Organizations

MultišŸ¤–Transformers's profile picture Cohere Labs Community's profile picture Hugging Face Discord Community's profile picture LangGraph UserGroup's profile picture Konkani LLM's profile picture Hugging Science's profile picture MCP-1st-Birthday's profile picture Hugging Face Skills's profile picture 25daysofagents's profile picture ML intern explorers's profile picture Mistral Hack-a-ton 2026's profile picture SAFETY AYA's profile picture Reuben's Multimodal Data Lab's profile picture Build Small Hackathon's profile picture