Suraj Srinivas
@Suuraj
Followers
1K
Following
8K
Media
38
Statuses
959
ml researcher / trying to understand why deep learning works
Sunnyvale, CA
Joined June 2009
RT @DimitrisPapail: Thinking about model generalization is quite painful. We observe empirically that models trained with SGD on cross-en….
0
57
0
RT @Michael_J_Black: Here's how my recent papers & reviews are going:. * To solve a vision problem today, the sensible thing is to leverage….
0
58
0
RT @alex_oesterling: ‼️🕚New paper alert with @ushabhalla_: Leveraging the Sequential Nature of Language for Interpretability ( https://t.co/….
0
8
0
Also, I'll be at ICML next week presenting this. Come say hi if you're around!.
0
0
3
It turns out that you can train your LLM by injecting benchmark eval data into your train data, and still have no effect on benchmark evals! . Accepted at @icmlconf. Joint work with @sbordt @valentynepii and Ulrike von Luxburg.
Have you ever wondered whether a few times of data contamination really lead to benchmark overfitting?🤔 Then our latest paper about the effect of data contamination on LLM evals might be for you!🚀. "How Much Can We Forget about Data Contamination?" (accepted at #ICML2025) shows
1
1
7
RT @ML_Theorist: Why does Chain of Thought prompting actually work?.@bohang_zhang will be talking about it today. Join us!. @Suuraj @tverven.
0
2
0
RT @GoodfireAI: We created a canvas that plugs into an image model’s brain. You can use it to generate images in real-time by painting wit….
0
98
0
we live in a world where "verification is easier than generation" is no longer true.
#NLProc.AI Co-Scientists 🤖 can generate ideas, but can they spot mistakes? (not yet! 🚫). In my recent paper, we introduce SPOT, a dataset of STEM manuscripts (math, materials science, chemistry, physics, etc), annotated with real errors. SOTA models like o3, gemini-2.5-pro
0
0
6
RT @ML_Theorist: ⏰⏰ Theory of Interpretable AI Seminar ⏰⏰.Chain-of-Thought: Why does explaining to LLMs using CoT prompting work?. Join us….
0
2
0
RT @norabelrose: data attribution is the most neglected thing in interpretability and people should join me in working on it.
0
4
0
RT @ML_Theorist: Curious about feature attribution? . SHAP & LIME treat features independently—but features interact!.Come hear how to "Dis….
0
1
0
RT @ML_Theorist: In April 2024, we launched the Theory of Interpretable XAI seminar, aiming to build a community—unsure if we’d even have e….
0
3
0
RT @ML_Theorist: ⏰⏰Theory of Interpretable AI Seminar ⏰⏰. Interested in Feature Attribution Explanations?. In two weeks, May 6, Gunnar Köni….
0
1
0
RT @ML_Theorist: Today in **two hours** @mirco_mutti will talk about interpretable bandits. Zoom link: @Suuraj @t….
0
1
0
RT @orthonormalist: DID I CRACK IT?. I think I figured out at least a chunk of the math. It's trade deficit divided by their exports. EU:….
0
2K
0