Danish Pruthi
@danish037
Followers
11K
Following
2K
Media
37
Statuses
508
Faculty at the Indian Institute of Science, Bangalore. PhD from @LTIatCMU.
Joined April 2014
Still can’t believe it. Our work on uncovering plagiarism in AI generated research received the outstanding paper award at ACL!!. Effort led by, and envisioned by the amazing @tarungupta360!.
54
31
563
Another paper was highlighted in SAC awards. 38: Silencing Empowerment, Allowing Bigotry: Auditing the Moderation of Hate Speech on Twitch. By Prarabdh Shukla, Wei Yin Chong, Yash Patel, Brennan Schaffner, Danish Pruthi, Arjun Bhagoji.
2
3
72
RT @cneuralnetwork: All That Glitters is Not Novel: Plagiarism in AI Generated Research by @tarungupta360 and @danish037 wins outstanding p….
0
6
0
RT @SafiKhan2k: Had an amazing time working on this with @danish037 and @im_mansigupta !. Catch @JankiNawale and me at our poster at ACL 2….
0
3
0
And lastly an audit of the state of hate-speech moderation @Twitch. Hall 4/5 on Monday at 11 AM. ArXiv:
0
0
1
FairI Tales: Evaluating fairness in Indian contexts with a focus on bias & stereotype. By Janki Atul Nawale, Mohammed Safi Ur Rahman Khan, Janani D, Mansi Gupta, Danish Pruthi, Mitesh M. Khapra. Hall 4/5 on Monday 11 AM. Details:
🚨🚨 Paper Alert!! 🚨🚨. Thrilled to announce our new paper, "FairI Tales: Evaluating fairness in Indian contexts with a focus on bias & stereotype” (FairI = "Fair"ness for "I"ndia, and yes, read as Fairy), is accepted at #ACL2025!. Paper 📄: Dataset 🤗:
0
0
4
All That Glitters is Not Novel: Plagiarism in AI Generated Research by Tarun Gupta, Danish Pruthi. Hall 4/5 on Tuesday at 4 PM. See this thread for details:
Remember this study about how LLM generated research ideas were rated to be more novel than expert-written ones? . We find a large fraction of such LLM generated proposals (≥ 24%) to be skillfully plagiarized, bypassing inbuilt plagiarism checks and unsuspecting experts. A 🧵.
2
0
4
I will be at #ACL2025 starting this Sunday. Will look forward to seeing old friends & making new ones. And yes, Indian Institute of Science is actively hiring faculty with strong NLP/AI background. No better time than now. We are presenting the following papers: 🧵.
6
1
81
As we have scaled our conferences up (somehow), there is absolutely no human-touch in the entire peer-review pipeline of most AI/NLP conferences. No wonder the quality has gone down the drain. Just feels like soul-sucking service. We need better models.
5
4
81
RT @pratyushmaini: At #ICML2025, I am super excited to introduce STAMP. This is a marriage b/w dataset inference & watermarking that finall….
0
14
0
Rarely does a book shape one's perspective as much as "The Socrates Express" did for me. The author, @Eric_Weiner, brings to life dead philosophers and what they stood for. It's been my go to book for months now. To be savored.
0
0
26
STAMP offers strong protection, can successfully detect membership of content that appears only once in the training data & constitutes < 0.001% of the total tokens. Work led by Saksham Rastogi, in collab w/ @pratyushmaini. Paper:
arxiv.org
Given how large parts of publicly available text are crawled to pretrain large language models (LLMs), data creators increasingly worry about the inclusion of their proprietary data for model...
0
1
13
STAMP creates multiple paraphrases of your content, through a watermarked model. Creators can then release one such paraphrased version, while keeping others private. Comparing model likelihoods on the private and public versions can effectively detect data membership. ⤵️.
1
0
7
At #ICML2025, introducing STAMP. A simple approach to verify whether your content (e.g., a dataset) is a part of the data used for training language models. ⤵️.
3
8
103
For ppl who are interested, the program details of the JTG/IEEE ITSoc Summer School:
ee.iitb.ac.in
JTG 2025 website
0
0
2
Will be at IIT Bombay on July 7-8, mostly to the rant about AI slop. DM if you are around and wanna chat.
5
1
45
RT @HarveenChadha: Wait until you find out how billing of engineers happen in Indian service based IT companies, soham will look like a sai….
0
21
0
RT @aagrawalAA: Danish Pruthi speaking about geographical disparities in language and image generation @vlms4all at #CVPR2025 (room: 104E).….
0
4
0