33 18 12

Leshem Choshen

borgr

https://ktilana.wixsite.com/leshem-choshen

AI & ML interests

Merging models, collaboratively improving pretraining, evaluation, understanding

Recent Activity

commented on an article 4 days ago

Cohere on Hugging Face Inference Providers 🔥

commented on an article 6 days ago

Cohere on Hugging Face Inference Providers 🔥

upvoted a paper 6 days ago

TextArena

View all activity

Organizations

borgr's activity

commented on Cohere on Hugging Face Inference Providers 🔥 4 days ago

Labelling, studying what people lack, learning about human reactions to various LM behavior etc.

commented on Cohere on Hugging Face Inference Providers 🔥 6 days ago

Great announcement!
Is there a way to opt in to share my data with the world when using the UI? Or get all my conversations with an API request (so others\we can build this opt in, in some hacky way)?

upvoted a paper 6 days ago

TextArena

Paper • 2504.11442 • Published 7 days ago • 27

authored a paper 12 days ago

Pretraining Language Models for Diachronic Linguistic Change Discovery

Paper • 2504.05523 • Published 15 days ago • 6

upvoted a paper 12 days ago

Pretraining Language Models for Diachronic Linguistic Change Discovery

Paper • 2504.05523 • Published 15 days ago • 6

commented a paper 12 days ago

Pretraining Language Models for Diachronic Linguistic Change Discovery

Paper • 2504.05523 • Published 15 days ago • 6 •

upvoted a paper 12 days ago

Fusing finetuned models for better pretraining

Paper • 2204.03044 • Published Apr 6, 2022 • 6

upvoted a paper 18 days ago

Scaling Analysis of Interleaved Speech-Text Language Models

Paper • 2504.02398 • Published 19 days ago • 27

upvoted an article 28 days ago

Article

FeeL: Making Multilingual LMs Better, One Feedback Loop at a Time

and 1 other •

28 days ago

• 12

published an article 28 days ago

Article

FeeL: Making Multilingual LMs Better, One Feedback Loop at a Time

and 1 other •

28 days ago

• 12

upvoted a paper about 1 month ago

Survey on Evaluation of LLM-based Agents

Paper • 2503.16416 • Published Mar 20 • 88

commented on From Chunks to Blocks: Accelerating Uploads and Downloads on the Hub about 1 month ago

That's great to know! Tx for sharing.

commented on From Chunks to Blocks: Accelerating Uploads and Downloads on the Hub about 1 month ago

Is all this data saved in raw fp? Or compressed somehow?

(yes I have an unhidden agenda of wondering due to zipNN lossless (open source) model compression and its relevance https://github.com/zipnn/zipnn)

upvoted a paper 2 months ago

Can this Model Also Recognize Dogs? Zero-Shot Model Search from Weights

Paper • 2502.09619 • Published Feb 13 • 35

upvoted a collection 3 months ago

Dicta-LM 2.0 Collection

Collection

9 items • Updated Apr 27, 2024 • 16

published 5 models 3 months ago