Pratyush Maini

pratyushmaini's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

liked a dataset 2 days ago

DatologyAI/SafeTulu

Viewer • Updated 14 days ago • 982k • 233 • 3

liked a Space 2 days ago

Running

Agents

7

Build Your Own SafeTulu

🛡

7

Filter safety-harming data out of Tulu 3 at any threshold

authored 9 papers 4 days ago

Follow the Gradient: Tracing Safety Degradation Back to Training Data

Paper • 2610.37481 • Published 11 days ago • 32

Natively Unlearnable Large Language Models

Paper • 2606.13873 • Published Jun 11

LLM Dataset Inference: Did you train on my dataset?

Paper • 2406.06443 • Published Jun 10, 2024

ÜberWeb: Insights from Multilingual Curation for a 20-Trillion-Token Dataset

Paper • 2602.15210 • Published Feb 25 • 1

The Finetuner's Fallacy: When to Pretrain with Your Finetuning Data

Paper • 2603.16177 • Published Mar 21

20/20 Vision Language Models: A Prescription for Better VLMs through Data Curation Alone

Paper • 2605.11405 • Published May 13

DatBench: Discriminative, Faithful, and Efficient VLM Evaluations

Paper • 2601.02316 • Published Jan 5 • 10

Luxical: High-Speed Lexical-Dense Text Embeddings

Paper • 2512.09015 • Published Dec 9, 2025

Safety Pretraining: Toward the Next Generation of Safe AI

Paper • 2504.16980 • Published Apr 23, 2025

upvoted a paper 4 days ago

Follow the Gradient: Tracing Safety Degradation Back to Training Data

Paper • 2610.37481 • Published 11 days ago • 32

submitted a paper to Daily Papers 4 days ago

Follow the Gradient: Tracing Safety Degradation Back to Training Data

Paper • 2610.37481 • Published 11 days ago • 32

updated 4 datasets 6 months ago

locuslab/fineweb_annotated

Viewer • Updated May 14 • 176M • 1.54k • 2

locuslab/refuseweb

Viewer • Updated May 14 • 1.65M • 848 • 4

locuslab/safeweb

Viewer • Updated May 14 • 14.8M • 1.48k • 5

locuslab/moral_education

Viewer • Updated May 14 • 2.81M • 991 • 4

liked 2 datasets 10 months ago

DatologyAI/DatBench

Viewer • Updated 5 days ago • 43.5k • 2.89k • 108

DatologyAI/DatBench-Full

Viewer • Updated 5 days ago • 152k • 1.12k • 77

upvoted a paper 10 months ago

DatBench: Discriminative, Faithful, and Efficient VLM Evaluations

Paper • 2601.02316 • Published Jan 5 • 10