Pradeep Dasigi
@pdasigi
Followers
1K
Following
685
Media
3
Statuses
446
Senior Research Scientist @allen_ai; #NLProc, Post-training for OLMo
Seattle, WA
Joined February 2009
RT @valentina__py: 💡Beyond math/code, instruction following with verifiable constraints is suitable to be learned with RLVR. But the set of….
0
92
0
RT @JesseDodge: @percyliang @AiEleuther nice! we also recently trained a set of models on 25 different pretraining corpora, each corpus hav….
0
7
0
RT @natolambert: A very exciting day for open-source AI! We're releasing our biggest open source model yet -- OLMo 2 32B -- and it beats th….
0
150
0
RT @srush_nlp: Really interesting paper on data selection. At this point just read everything Hamish writes.
0
18
0
How to curate instruction tuning datasets while targeting specific skills? This is a common question developers face while post-training LMs. In this work led by @hamishivi, we found that simple embedding based methods scale much better than fancier computationally intensive.
How well do data-selection methods work for instruction-tuning at scale?. Turns out, when you look at large, varied data pools, lots of recent methods lag behind simple baselines, and a simple embedding-based method (RDS) does best!. More below ⬇️ (1/8)
0
0
12
RT @HannaHajishirzi: Excited to drive innovation and push the boundaries of open, scientific AI research & development! 🚀 Join us at @allen….
0
15
0
RT @hamishivi: One additional thing in the updated Tulu 3 paper that I'd like to highlight is that @pdasigi went back and re-evaluated our….
0
1
0
Here's a significant update to Tülu 3: we scaled up the post-training recipe to Llama 3.1 405B. Tülu 3 405B beats Llama's 405B instruct model and also Deepseek V3. You can now access the model and the entire post-training pipeline. Huge shoutout to @hamishivi and @vwxyzjn who.
Here is Tülu 3 405B 🐫 our open-source post-training model that surpasses the performance of DeepSeek-V3! The last member of the Tülu 3 family demonstrates that our recipe, which includes Reinforcement Learning from Verifiable Rewards (RVLR) scales to 405B - with performance on
0
0
12
RT @faeze_brh: Just arrived in 🇨🇦 to attend NeurIPS 2024!. Excited to connect and chat about AI reliability and safety, resource-efficient….
0
3
0
RT @saumyamalik44: I'm having a great time as a PYI at Ai2! Consider applying for this great program :).
0
10
0
Our team at Ai2 (OLMo) is looking for a predoctoral researcher. You get to work on exciting research in building open LMs while preparing for a PhD. Apply here:
job-boards.greenhouse.io
0
0
2
RT @interconnectsai: OpenAI's o1 using "search" was a PSYOP.How to understand OpenAI's o1 models as really just one wacky, wonderful, long….
www.interconnects.ai
How to understand OpenAI's o1 models as really just one wacky, wonderful, long chain of thought
0
22
0
RT @valentina__py: Applying our Tülu 3 recipe to OLMo 2 resulted in OLMo 2 Instruct, which is competitive with even the best open-weight mo….
0
1
0