## Recapping 2024 in Synthetic Data/Smol Models! We are very honored to have @LoubnaBenAllal1 recap and pick all the best papers in: Synthetic Data and Smol Models this year! Timestamps [00:00:05] Loubna Intro
[00:00:33] The Rise of Synthetic Data Everywhere
[00:02:57]
LLMS
-
2024 Synthetic Data and Small Models: Best Papers Recap
By
–
-
Qwen Model License Clarification Apache 2.0
By
–
Qwen is just Apache 2.0 tho – messaged them to rectify
-
Converting Models to Llama.cpp Format Guide
By
–
Pretty much yes! Just need to convert to llama.cpp format
-
Clarifying reported model ensemble scores on validation vs semi-private set
By
–
Sorry, but I think you’re misinterpreting the paper: The 62% is an ensemble of TTT + BARC on the public validation set and doesn’t imply any score on the harder semi-private v1. It shouldn’t be on that plot.
-
Chain of Thought and RL in Extended AI Reasoning Analysis
By
–
the idea of analyzing things over a long derivation is about 60 years old, and will be part of the final answer; the particular way that o* does it (presumably via Chain of thought and RL but maybe other mechanisms) has not been transparently explained, and may or may not play a
-
AI Landscape Discussion: Reasoning, Hardware, Startups and Policy
By
–
daniel and i spent 2 hours together yesterday discussing ai over the last year – reasoning, bio, chips, startups, vibe shifts, policy, and ofc wassup @airstreet 🙂
— Nathan Benaich (@nathanbenaich) 24 décembre 2024
episode dropping soon on @gradientpub 💙 https://t.co/jZhfxORjIHdaniel and i spent 2 hours together yesterday discussing ai over the last year – reasoning, bio, chips, startups, vibe shifts, policy, and ofc wassup @airstreet 🙂 episode dropping soon on @gradientpub
-
Llama 3.3 70B: Advanced Context Lengths for AI Agent Development
By
–
Llama 3.3 70B is the gift that keeps on giving for all the #devs out there🦙🎁 @AIatMeta
— SambaNova (@SambaNovaAI) 24 décembre 2024
You can explore context lengths of 64K tokens & build #Agents with other models — all on SambaNova Cloud!
Need higher rate limits? Just ask! 💭
Try it now👇 https://t.co/Nq7AnfsVMP pic.twitter.com/nBMYdnVOQPLlama 3.3 70B is the gift that keeps on giving for all the #devs out there @AIatMeta You can explore context lengths of 64K tokens & build #Agents with other models — all on SambaNova Cloud! Need higher rate limits? Just ask! Try it now https://
cloud.sambanova.ai -
O3 unreliable across domains, better AI technology will emerge
By
–
Aside from details around ARC & other benchmarks, o3 won’t turn out to be reliable across domains, and it will in hindsight seem wildly inefficient. MUCH better technology will emerge, after the LLM hype subsides. (LLMs may still exist, but will be just one tool in a larger
-
Qwen Releases QVQ-72B-Preview: Advanced Multimodal AI Model
By
–
We're so unfathomably back! https://
huggingface.co/Qwen/QVQ-72B-P
review
…