Aaditya Singh
@Aaditya6284
Followers
673
Following
1K
Statuses
355
Doing a PhD @GatsbyUCL with @SaxeLab, @FelixHill84 on learning dynamics, ICL, LLMs. Prev. at: @GoogleDeepMind, @AIatMeta (LLaMa 3), @MIT. https://t.co/ZOmBWCvbIK
London, UK
Joined May 2022
Math reasoning benchmarks keep getting saturatedโฆ Excited to introduce HARD-Math: Human-Annotated Reasoning Dataset for Math. Consisting of 4,780 short answer problems, based on the AHSME, AMC, & AIME contests, HARD-Math still poses a challenge for frontier LLMs. Read on ๐โฌ
1
24
154
RT @alexwei_: o3 learns its own tricks through more RL, such as testing its code on random test cases against a brute forceโa technique I uโฆ
0
14
0
RT @RamanujanVivek: Happy to (belatedly) share our recent work introducing Causally Regularized Tokenization ๐บ, matching LlamaGen-3B generaโฆ
0
26
0
RT @__nmca__: Epoch AI are going to publish more details, but on the OpenAI side for those interested: we did not use FrontierMath data toโฆ
0
36
0
RT @GatsbyUCL: ๐ข Apply now for the 7-week #Mathematics #SummerSchool in London! You will develop mathematics skills and intuition necessarโฆ
0
53
0
RT @scychan_brains: Devastatingly, we have lost a bright light in our field. Felix Hill was not only a deeply insightful thinker -- he wasโฆ
0
93
0
Thank you for investing in me, a master's student with no papers. Thank you for believing in me, through health ups and downs. Thank you for transforming my life. Thank you for always bringing people together. May you rest in peace ๐ญโค๏ธ you'll always be a hero to me.
1
0
17
RT @allisontam_: rest in peace Felix. I hope that you can see all the positive impact youโve had in our lives. to those reading, you are nโฆ
0
3
0
RT @DavidDingAI: Iโm incredibly sad to hear about the passing of @FelixHill84. Iโve had the pleasure of working closely with Felix for my fโฆ
0
1
0
RT @AndrewLampinen: @FelixHill84 was such an incredible mentor โ and occasional cold water swimming partner โ to me. He's a huge part of whโฆ
0
10
0
RT @Yihe__Deng: I want to share a new blog, in collaboration with @YUYANG_UCLA, on ๐๐ฉ๐ฎ๐ซ๐ข๐จ๐ฎ๐ฌ ๐๐จ๐ซ๐ซ๐๐ฅ๐๐ญ๐ข๐จ๐ง, ๐๐ก๐จ๐ซ๐ญ๐๐ฎ๐ญ ๐๐๐๐ซ๐ง๐ข๐ง๐ , ๐๐ง๐ ๐๐๐ฐ๐๐ซ๐ ๐๐๐๐คโฆ
0
14
0
Was super cool to be a part of this event!
It was incredible to host our first panel event with @SWC_Neuro on "Will computational models and machine learning ever be able to replicate the full complexity of the brain?" ๐ง ๐ป๐ฌ If you didnโt get a chance to attend, have a watch now! ๐
0
1
12
RT @sama: if you are a safety researcher, please consider applying to help test o3-mini and o3. excited to get these out for general availaโฆ
0
587
0
RT @__nmca__: o3 represents enormous progress in general-domain reasoning with RL โ excited that we were able to announce some results todaโฆ
0
381
0
RT @shengjia_zhao: Excited to train o3-mini with @ren_hongyu @_kevinlu and others, a blindingly fast model with amazing reasoning / code /โฆ
0
44
0
RT @lukaszkaiser: @apples_jimmy On the 12th day of shipmas Santa brought to me... AGI in a model called ooo3
0
9
0
RT @fchollet: Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it reโฆ
0
2K
0