Third year CS PhD candidate at Princeton University (@princeton_nlp @PrincetonPLI), previously CS undergrad at IIT Bombay
Princeton, NJ
Based in United States
- Tweets 135
- Following 603
- Followers 1,042
- Likes 533
Pinned Tweet
The most profound paper of my PhD so far. We truly did something special to make LMs play/explain chess, needing both architectural & algorithmic innovation (natural-language analogue of the Alphazero algorithm) + applicable to many other domains. Please read on and share!
1/10
55
168
1,522
103,670
Happening now at Grand Ballroom Poster #133! Stop by!
Language models that think, chat better.
We used longCoT (w/ reward model) for RLHF instead of math, and it just works. Llama-3.1-8B-Instruct + 14K ex beats GPT-4o (!) on chat & creative writing, & even Claude-3.7-Sonnet (thinking) on AlpacaEval2 and WildBench!
Read on. 🧵
1/8
14
Adithya Bhaskar retweeted
🚨SciConBench accepted at #NeurIPS2026!
Paper: arxiv.org/abs/2606.11337
Can AI *synthesize* scientific conclusions in health? We find that, on medical conclusions, we are far from it!
Check out our live dashboard: sciconbench.cs.princeton.edu…
w/@hayounggjung and @korolova
3
17
52
4,618
Adithya Bhaskar retweeted
Super proud of what the team has achieved @AdithyaNLP @jeff_cheng_77 (they are both great researchers and chess players)
A 4B chess LM that plays chess at typical grandmaster level and explains its moves (!)
I will discuss this work at Berkeley in 30 minutes.
Also, I am around this week at @COLM_conf!
The most profound paper of my PhD so far. We truly did something special to make LMs play/explain chess, needing both architectural & algorithmic innovation (natural-language analogue of the Alphazero algorithm) + applicable to many other domains. Please read on and share!
1/10
3
16
228
13,874
We are both at COLM, come talk to us if you are interested in chess / interpretation of silent experts!
‼️Have you ever used a chess engine and wished it could explain its moves and evaluations?‼️
Introducing QUEEN, a 4B-parameter chess-language model that plays at the level approaching a typical Grandmaster and explains its moves in natural language.
13
1,247
The most profound paper of my PhD so far. We truly did something special to make LMs play/explain chess, needing both architectural & algorithmic innovation (natural-language analogue of the Alphazero algorithm) + applicable to many other domains. Please read on and share!
1/10
55
168
1,522
103,670
Here is an explanation by that "templatic" model - not so templatic after three rounds of self-improvement after all!
9/10
1
1
51
4,456
Huge thanks to my amazing co-author and chess buddy @jeff_cheng_77, and our advisor @danqi_chen! You can find QUEEN & more ablations/analyses at
- Paper: arxiv.org/abs/2610.03695
- Website: queen-project.github.io/
- HF: huggingface.co/papers/2610.0…
10/10
8
5
100
4,790
I will present this paper as a poster at COLM 2026 (Grand Ballroom #133, Franciscan, Poster #1199, F-4-133)!
HMU for research/coffee chats or with suggestions for good coffee shops! I've been working on making LLMs learn from non-language data, and I'll have more to say soon.
Language models that think, chat better.
We used longCoT (w/ reward model) for RLHF instead of math, and it just works. Llama-3.1-8B-Instruct + 14K ex beats GPT-4o (!) on chat & creative writing, & even Claude-3.7-Sonnet (thinking) on AlpacaEval2 and WildBench!
Read on. 🧵
1/8
2
3
29
1,548
Adithya Bhaskar retweeted
Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI.
Perhaps counterintuitively, I think it's a great time to do so.
I wrote up some thoughts on this here: web.mit.edu/phillipi/www/wri…
92
582
3,848
1,732,055
THE PURPOSE OF MATHEMATICS IS NOT TO PROVE THINGS OR TO SOLVE PROBLEMS: it is to gain a better understanding of the mathematical universe and how it works (either per se → pure math, or to use this understanding elsewhere → applied math). •3/38
2
11
101
6,483
Adithya Bhaskar retweeted
thinking of this heartbreaking essay written by a mathematician last week, before this most recent news dropped
An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. openai.com/index/ten-advance…
260
1,329
17,725
2,639,287
Adithya Bhaskar retweeted
My work in 2023: think deeply about technical topic, build the perfect codebase from scratch, know it by heart
My work in 2026: chatting with five dumb people at the same time, they've been stuck Whirliging... for fifteen minutes, not sure what to do so i might spin up a sixth
13
14
541
21,638