Top Tweets for #Neurips26
[1/n] Iโm thrilled to share that the ROTATE paper, which casts learning generalizable coordination behaviors as an open-ended learning problem, has been accepted to #NeurIPS26 as a spotlight ๐! See you in Sydney!
Collaborators: @ArrasyRahman, Benjamin Nativi, Johnny Liu, @cuijiaxun, @yoonchangsung, @PeterStone_TX.
Paper: https://t.co/DrD8KEt8WT
Code: https://t.co/QaACEHqE3p
For those of you who might have seen earlier versions of the paper, in the most recent version, we made substantial advances by characterizing the self-sabotage problem as per-state incompetence, and viewing ROTATE as placing a competence constraint on learning, therefore improving our understanding of why ROTATE works. Details in the thread below.
![CarolineWang98's tweet photo. [1/n] Iโm thrilled to share that the ROTATE paper, which casts learning generalizable coordination behaviors as an open-ended learning problem, has been accepted to #NeurIPS26 as a spotlight ๐! See you in Sydney!
Collaborators: @ArrasyRahman, Benjamin Nativi, Johnny Liu, @cuijiaxun, @yoonchangsung, @PeterStone_TX.
Paper: https://t.co/DrD8KEt8WT
Code: https://t.co/QaACEHqE3p
For those of you who might have seen earlier versions of the paper, in the most recent version, we made substantial advances by characterizing the self-sabotage problem as per-state incompetence, and viewing ROTATE as placing a competence constraint on learning, therefore improving our understanding of why ROTATE works. Details in the thread below.](https://pbs.twimg.com/media/HU8C3XUWkAAGcWX.jpg)
@nickcammarata Harness for automatic neuron labelling, take a look at this https://t.co/TxaNFVg2R6 #NeurIPS26 by @ccaiwut
Can we replace individual neural network weights with human-readable Python code? Sorry to be late... but I'm super thrilled to share that our paper was accepted to #NeurIPS26.
"Individual Parameters in Weight-Sparse Transformers Appear Interpretable" (w/ @sheimersheim!!)
What do we actually know about how Llama 3 behaves?
Modelpedia collects 1,000+ model findings from ICLR, linked to models, datasets and methods. Most of them rest on correlation, not intervention. #Neurips26 MetaScience Workshop
https://t.co/CevmfdRvoy
๐ https://t.co/RN5A9VLRAh
ScAI is heading to #NeurIPS26 with 15 papers
Dialogue Systems, Reinforcement Learning, LLMs, graph ML, Materials, Scientific discovery, Recommendation System, AI for Science & Healthcare a pretty wide spread this year.
Congrats to all the authors.
Let the kids @barbera_cesare @samubortolotti @NDebole60493 cook! Super glad to have been part of our new papers accepted @ #NeurIPS26 (with others amazing coauthors) on local calibration, conformal prediction for NeSy and Gaussian Processes for CBMs! More info to come out soon!
SML goes to #NeurIPS26 with four new works on link prediction, local calibration, conformal prediction for Nesy and Gaussian Processes for CBMs!
If you want to know more about them, links to the arXiv version below ๐ and more info to come out soon!
SML goes to #NeurIPS26 with four new works on link prediction, local calibration, conformal prediction for Nesy and Gaussian Processes for CBMs!
If you want to know more about them, links to the arXiv version below ๐ and more info to come out soon!
Here is the story motivating our IAB workshop: Several months ago I was reading the agent literature. The basic agent architecture is now roughly settled, and papers describe a few primary behaviors, mostly planning and reasoning. But what are other behaviors?
Drift Q-Learning just got into #Neurips26 ! ๐ฆ๐บ
I guess see you in Australia!!
Project Page: https://t.co/a003Qo4m9h
Diffusion model โ
Generated item โ
Successful item โ
โก๏ธ Now: Which training data was influential?
โก๏ธ Which are the items that, without them, such generation would not have been there?
This is the counterfactual question our latest paper tries to answer more accurately.
1/5
This paper is accepted by #NeurIPS26.
The title of this work has been changed to "PORTool: Importance-Aware Policy Optimization with Rewarded Tree for Multi-Tool-Integrated Reasoning". More updates of this work will come shortly.
Paper: https://t.co/Rt21ee8DL6
Excited to share our latest work at Apple: PORTool: Tool-Use LLM Training with Rewarded Tree ๐
In this work, we reinforce LLMs for multi-step multi-tool use when only the final outcome is available.
Paper: https://t.co/Rt21ee8DL6
#LLM #ToolUse #AIResearch ๐งต1/n
Today we get the NeurIPS 2026 results ๐
For reference, here are the official NeurIPS 2025 statistics on acceptance rates by average reviewer score.
#NeurIPS2026 #NeurIPS26 #NeurIPS2025 #NeurIPS25
Last Seen Hashtags on Sotwe
minichat
Seen from Chile
beylikduzutrave๏ฌ
i
Seen from Turkey
teenageers
Seen from Turkey
ูุนูุฑููู
Seen from Saudi Arabia
เธเธญเธเนเธเนเธ
Seen from Thailand
๋ฉ์ญ
Seen from Korea
unexpected() filter:videos
Seen from United Kingdom
anxiouspandaa()()()()()filter:videos
Seen from Peru
kฤฑrลehiraktif
Seen from Turkey
Most Popular Users
Elon Musk 
@elonmusk
241.5M followers
Barack Obama
@barackobama
118.8M followers
Cristiano Ronaldo
@cristiano
114.6M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.3M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
90.1M followers
Taylor Swift
@taylorswift13
84M followers
Lady Gaga
@ladygaga
75.5M followers
Virat Kohli
@imvkohli
73.5M followers
Kim Kardashian
@kimkardashian
70.9M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.5M followers
Bill Gates
@billgates
65.3M followers
Selena Gomez
@selenagomez
63.1M followers
The Ellen Show
@theellenshow
62.2M followers
CNN
@cnn
61.7M followers
X
@x
60.7M followers