Somesh
@someshsingh22
RA @adobe | CS PhD @UB & @IIITD CS 22 #bitsg |
3/3 Accepts and all positive reviews💗Excited to share the acceptance of our work on Overstability and Oversensitivity in Automatic Essay Scoring (AES) systems in the #journal of Dialogues and Discourse! Camera-ready version is arriving soon. arxiv.org/pdf/2109.11728… :🧵
This, Chained Agents work because we segregate the correct context in the window where LLMs actually understand them. Multi-Agent frameworks combine this with system roles (because of SFT) and tools. This transforms everything into a neat POMDP!
+1 for "context engineering" over "prompt engineering". People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window…
We've been using uv a few months now and I've never felt better. I have more energy. My skin is clearer. My eye sight has improved.
There are traditionally two types of research: problem-driven research and method-driven research. As we’ve seen with large language models and now AlphaEvolve, it should be very clear now that total method-driven research is a huge opportunity. Problem-driven research is nice…
torch.compile on QLora, huh that- Using hugging face
We made 5 challenges and if you score 47 points we'll offer you $500K/year + equity to join us at 🦥@UnslothAI! No experience or PhD needed. $400K - $500K/yr: Founding Engineer (47 points) $250K - $300K/yr: ML Engineer (32 points) Challenges: 1. Convert nf4 / BnB 4bit to…
LLMs will solve this in the coming week 😉 Never if you show an incomplete puzzle, either send the PGN or highlight the last move tho 🙃
If and when @OpenAI , @deepseek_ai , @AnthropicAI and @GoogleGemi can solve this chess problem, I’ll believe LLMs can reason. White to checkmate in 1 move @GaryMarcus
Excited to share our new work on test-time alignment! We introduce HyRe, a fast way to adapt large models (like LLM reward models) to new user preferences without extra training. Paper: arxiv.org/abs/2412.08812
Don’t race. Don’t catch up. Don’t play the game. Instead, do rigorous science. Do controlled experiments. Formulate clear hypothesis. Carefully examine alternative hypothesis. Rule out confounders. Listen to the physics of LLM tutorial 10 times and recite every single word of it.…
The gap between open-sourced models and closed-source models is getting larger and larger. What should academia do to catch up?
It's really interesting to see the "human difficulty" of sharp positions combined with time pressure still evade most known statistical methods, Na1 was a genius move in hindsight. GG @DGukesh. Looks like SOTA engines still have a great scope ;) #DingGukesh
Extremely sharp and dangerous position with little time is a formula for inaccuracies and blunders. #DingGukesh
@sbhatia_ is attending @emnlpmeeting Set up some time to talk to him to know more about the job and the lab #NLProc #EMNLP2024 #EMNLP
We are hiring a Research Scientist in the field of Behavioral Sciences and Machine Learning at Adobe MDSR Labs. Requisites: PhD or equivalent experience Job Description: acrobat.adobe.com/id/urn:aaid:sc…
We are hiring a Research Scientist in the field of Behavioral Sciences and Machine Learning at Adobe MDSR Labs. Requisites: PhD or equivalent experience Job Description: acrobat.adobe.com/id/urn:aaid:sc…
Internship Opportunity at Adobe MDSR Labs We are seeking PhD and MS Research Interns in the fields of NLP, Computer vision, and Behavioral Sciences for Summer 2025 and 6 months. Send your resume to [email protected] Job Posting: acrobat.adobe.com/id/urn:aaid:sc…
Cool
@someshsingh22 I’ll send you IMPORTANT ELECTION UPDATES for North Carolina. Make sure you are ready to VOTE FOR DONALD J. TRUMP by November 5th. Reply #stop to opt-out.
SpiceJet Standard Time: 45 mins after expected @flyspicejet