Neemesh

492 posts

Neemesh banner
Neemesh

Neemesh

@stealth_py

reasoning + css @ smu, mpi-is || prev. @lcs2lab, ugrip'23 @mbzuai, ug @iiitdelhi || i love soulslike games but can also be found on the rift

Singapore Katılım Ağustos 2016
424 Takip Edilen208 Takipçiler
Neemesh
Neemesh@stealth_py·
oh well this was random
Neemesh tweet media
English
8
0
27
1.5K
Neemesh retweetledi
Aflah 🍉🕊️
Aflah 🍉🕊️@Aflah02101·
Really happy to share that three of our papers have been accepted to ICLR 2026! 1️⃣In Agents We Trust, but Who Do Agents Trust? Latent Preferences Steer LLM Generations Preview: openreview.net/pdf?id=yTUNl6j…
English
1
2
27
2K
Neemesh
Neemesh@stealth_py·
@Aflah02101 now play hollow knight, ds3, elden ring, sekiro, OR league of legends the best game of all time
English
1
0
1
39
Neemesh
Neemesh@stealth_py·
@yukti_x It really is 😭 also the goofy songs from clash royale reels though not sure if your fyp is as bricked
English
0
0
0
59
yuyu
yuyu@yukti_x·
we are charlieeee kirkkkkk is such an ear worm
English
1
0
7
533
Neemesh
Neemesh@stealth_py·
@Aflah02101 @kylelostat i dont think they remove repeated docs, but repeated boilerplate text under each doc for substring dedupe. although not sure how optimal 80% span repetition cutoff is, maybe that 20% can be useful but i would assume most of that 20% is dispersed in the span and not contiguous
English
1
0
0
15
Aflah 🍉🕊️
Aflah 🍉🕊️@Aflah02101·
@kylelostat .. so it is surprising to me that this doesn't lead to loss of a lot of useful data? I am probably misunderstanding something here
English
1
0
0
22
Kyle Lo
Kyle Lo@kylelostat·
olmo 3 paper finally on arxiv 🫡 thx to our teammates esp folks who chased additional baselines thx to arxiv-latex-cleaner and overleaf feature for chasing latex bugs thx for all the helpful discussions after our Nov release, best part of open science is progressing together!
Kyle Lo tweet media
English
11
98
458
52.7K
Neemesh retweetledi
Siddhant Agarwal
Siddhant Agarwal@Siddhant_62·
Our paper “MAMA-Memeia! Multi-Aspect Multi-Agent Collaboration for Depressive Symptoms Identification in Memes” is accepted to AAAI-26! 🎉 arxiv and code coming soon! I’m at NeurIPS 2025 (Dec 1-7) & looking for Summer 2026 internships, happy to connect! #AAAI26 #NeurIPS2025
English
0
2
6
685
Mathieu
Mathieu@miniapeur·
Mathieu tweet media
ZXX
5
18
344
13K
Neemesh retweetledi
LoL Esports
LoL Esports@lolesports·
THE BACK-TO-BACK CHAMPIONS RETURN TO THE SEMIFINALS.
LoL Esports tweet media
English
135
4.2K
17.1K
644.8K
Neemesh retweetledi
Johnny Tian-Zheng Wei
Johnny Tian-Zheng Wei@johntzwei·
Announcing 🔭✨Hubble, a suite of open-source LLMs to advance the study of memorization! Pretrained models up to 8B params, with controlled insertion of texts (e.g., book passages, biographies, test sets, and more!) designed to emulate key memorization risks 🧵
Johnny Tian-Zheng Wei tweet media
English
2
40
130
47.8K
Aflah 🍉🕊️
Aflah 🍉🕊️@Aflah02101·
Finished two of my favorite research projects this ICLR. It's funny both projects lie on opposite ends of the LLM development/usage process. One does heavy pretraining while the other looks at the agentic layer Hopefully preprints soon!
English
1
0
8
246
Neemesh
Neemesh@stealth_py·
@Aflah02101 @willccbb Ayooo 😭😭 (can I get some merch tho not sure if I qualify but I am chronically online)
English
0
0
1
13
Neemesh
Neemesh@stealth_py·
@Aflah02101 💯 If you yourself are not aligned to basic human morals and ethics then you should neither be working on "aligning" role-playing models to them nor should you be left anywhere close to children (if you think this is acceptable).
English
0
0
2
42
Aflah 🍉🕊️
Aflah 🍉🕊️@Aflah02101·
@stealth_py I've come to realize that there are close to no good role models in tech. You may admire people for their technical expertise but if this is the kind of crap they're gonna use their expertise for then they're not worth looking up to
English
1
0
1
62
Neemesh
Neemesh@stealth_py·
Very excited to be a part of this very important and amazing line of work!
Zhijing Jin@ZhijingJin

Really enjoyed the @NLP4PosImpact Workshop at @ACLmeeting #ACL2025NLP! I am glad to give the first version of our "LLMs and Democracy Defense" talk ⬇️ youtu.be/NpGpq1WKcuI We will also present at the upcoming NLP4Democracy Workshop at #COLM2025 in Montreal. Thx to all organizers for your tremendous work creating a platform for this important line of work. And thank you to all our brilliant students, and my long-term collaborator and mentor @radamihalcea on numerous inspirations❤️

English
0
0
3
243
Neemesh
Neemesh@stealth_py·
If you're interested for more details, kindly check out our work at arxiv.org/abs/2503.05280 Or, if you are attending ACL 2025, you are very welcome to drop by our poster! This work will also be presented as a poster at the NLP4PI workshop on July 31 in Exhibit Hall X5, 11-12 PM.
English
1
0
1
125
Neemesh
Neemesh@stealth_py·
Our work on censorship decisions will be presented during Session 12 from 11-12:30 PM in Hall 4/5 at ACL. Unfortunately, I won't be able to attend ACL due to VISA issues :(( However, Zhijing will be there to present our interesting findings in my place! Be sure to drop by!! ^_^
Neemesh tweet media
Zhijing Jin@ZhijingJin

🎉 Excited to have 4 papers at #ACL2025! We: (1) Train LLMs to reveal Twitter censorship in 5 countries (2) Probe+steer reasoning vs memorization in LLMs (3) Build a capable coding agent (4) Improve MT on AI research terms 🫶 Thanks to all collaborators! #ACL2025NLP @aclmeeting

English
1
5
16
1.4K

Keşfet

@Aflah02101 @yukti_x @kylelostat @miniapeur @willccbb @elonmusk @BarackObama @taylorswift13