archit's picture

archit PRO

archit11

·

archit-spec

AI & ML interests

small language models, looking for work please reachout [email protected]

Recent Activity

liked a dataset 2 days ago

CK0607/polymarket_10000

new activity 2 days ago

canopylabs/3b-hi-pretrain-research_release:Does the training data contain hinglish too?

new activity 2 days ago

canopylabs/3b-hi-pretrain-research_release:Does the training data contain hinglish too?

View all activity

Organizations

upvoted an article 3 days ago

Article

State of open video generation models in Diffusers

By

and 2 others •

Jan 27

• 54

upvoted 2 articles 9 days ago

Article

How Long Prompts Block Other Requests - Optimizing LLM Performance

By

•

15 days ago

• 2

Article

Prefill and Decode for Concurrent Requests - Optimizing LLM Performance

By

•

Apr 16

• 18

upvoted a paper 17 days ago

Reinforcement Pre-Training

Paper • 2506.08007 • Published 17 days ago • 232

upvoted a paper 19 days ago

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning

Paper • 2506.01939 • Published 24 days ago • 162

upvoted an article 3 months ago

Article

Enabling Long Context Training with Sequence Parallelism in Axolotl

By

and 1 other •

Apr 4

• 9

upvoted an article 4 months ago

Article

SigLIP 2: A better multilingual vision language encoder

By

and 2 others •

Feb 21

• 169

upvoted an article 5 months ago

Article

The case for specialized pre-training: ultra-fast foundation models for dedicated tasks

By

•

Aug 4, 2024

• 30

upvoted 3 collections 5 months ago

Scotch & SOTA 🥃 Pt. 7: Human Feedback Datasets 🫣

The elusive “human” feedback • 1 item • Updated Sep 13, 2023 • 1

Scotch & SOTA 🥃 Pt. 6: Dialogue Tuning Datasets 💬

Conversations, turn-based dialog, and things that can be turned into that. • 4 items • Updated Sep 13, 2023 • 1

Scotch & SOTA 🥃 Pt. 5: Instruction Tuning Datasets 👩‍🏫

Question & answer, task completion, general SFT and otherwise finetuney data. • 7 items • Updated Sep 13, 2023 • 1

upvoted 2 articles 5 months ago

Article

How to deploy and fine-tune DeepSeek models on AWS

By

and 2 others •

Jan 30

• 52

Article

Can we create pedagogically valuable multi-turn synthetic datasets from Cosmopedia?

By

•

May 7, 2024

• 8

upvoted a collection 5 months ago

Deepseek Papers

Deepseek papers collection • 24 items • Updated 3 days ago • 259

upvoted an article 5 months ago

Article

Train 400x faster Static Embedding Models with Sentence Transformers

By

•

Jan 15

• 195

upvoted 4 collections 7 months ago

Reasoning

152 items • Updated 29 days ago • 30

🤖 Agents

21 items • Updated Dec 31, 2024 • 160

Tulu 3 Datasets

All datasets released with Tulu 3 -- state of the art open post-training recipes. • 33 items • Updated Apr 30 • 86

PixMo

A set of vision-language datasets built by Ai2 and used to train the Molmo family of models. Read more at https://molmo.allenai.org/blog • 10 items • Updated Apr 30 • 72

upvoted a paper 7 months ago

Gemma 2: Improving Open Language Models at a Practical Size

Paper • 2408.00118 • Published Jul 31, 2024 • 78