Shantanu PRO

ShantanuT01

https://shantanuthorat.io

ShantanuT01
shantanu-thorat-cambridge

AI & ML interests

None yet

Recent Activity

authored a paper 11 days ago

Team DACTYL at PAN 2026: Bayesian Data Mixing and Empirical X-risk Minimization for AI-text Detection

updated a collection 11 days ago

PAN@CLEF 2026

published a model about 2 months ago

ShantanuT01/gradient-ai-text-detector

View all activity

Organizations

None yet

ShantanuT01 's collections 6

PAN@CLEF 2026

Models submitted to PAN CLEF 2026 Task 1.

ShantanuT01/vanguard-ai-text-detector

Text Classification • 0.4B • Updated Jul 18 • 2.76k
ShantanuT01/gradient-ai-text-detector

Text Classification • 0.4B • Updated Jul 18 • 366
Team DACTYL at PAN 2026: Bayesian Data Mixing and Empirical X-risk Minimization for AI-text Detection

Paper • 2607.17382 • Published Jul 19 • 1

DACTYL

Contains classifiers, fine-tuned SLMs, and datasets used for the Diverse Adversarial Corpus of Texts Yielded from Large language models.

DACTYL Classifiers

Collection

Trained AI-generated text classifiers. Pretrained means using binary cross entropy loss, finetuned refers to deep X-risk optimized classifiers. • 10 items • Updated Sep 24, 2025
DACTYL Finetuned SLMs

Collection

These are models that were continued pretrained on a domain-specific corpus. • 18 items • Updated Mar 2
ShantanuT01/DACTYL

Viewer • Updated Sep 24, 2025 • 655k • 73
ShantanuT01/DACTYL-Pretraining

Viewer • Updated Sep 25, 2025 • 416k • 47

DACTYL Finetuned SLMs

These are models that were continued pretrained on a domain-specific corpus.

ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-training

Text Generation • 1B • Updated Apr 29, 2025 • 4
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-validation

Text Generation • 1B • Updated Apr 29, 2025 • 1
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-testing

Text Generation • 1B • Updated Apr 29, 2025 • 3
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-reviews-validation

Text Generation • 1B • Updated Apr 29, 2025 • 4

RIP Dataset

ShantanuT01/RIP-Dataset

Viewer • Updated Oct 23, 2024 • 19k • 25
Which LLMs are Difficult to Detect? A Detailed Analysis of Potential Factors Contributing to Difficulties in LLM Text Detection

Paper • 2410.14875 • Published Oct 18, 2024

DACTYL Classifiers

Trained AI-generated text classifiers. Pretrained means using binary cross entropy loss, finetuned refers to deep X-risk optimized classifiers.

ShantanuT01/dactyl-deberta-v3-large-pretrained

Text Classification • 0.4B • Updated Sep 24, 2025 • 4
ShantanuT01/dactyl-ai-text-detector

Text Classification • 0.4B • Updated Mar 4 • 1.11k
ShantanuT01/dactyl-modernbert-base-pretrained

Text Classification • 0.1B • Updated Sep 24, 2025 • 2
ShantanuT01/dactyl-modernbert-base-finetuned

Text Classification • 0.1B • Updated Sep 24, 2025 • 12

R255

R255 Project Models and Datasets

ShantanuT01/Llama-3.2-1B-Instruct-yelp-finetuned

1B • Updated Mar 27, 2025 • 5
ShantanuT01/Llama-3.2-1B-Instruct-abliterated-yelp-finetuned

Text Generation • 1B • Updated Mar 28, 2025 • 13 •
ShantanuT01/R255-Finetuning-Datasets

Preview • Updated Mar 28, 2025 • 23
ShantanuT01/Llama-3.2-1B-Instruct-abliterated-wp-finetuned

1B • Updated Mar 28, 2025 • 8

PAN@CLEF 2026

Models submitted to PAN CLEF 2026 Task 1.

ShantanuT01/vanguard-ai-text-detector

Text Classification • 0.4B • Updated Jul 18 • 2.76k
ShantanuT01/gradient-ai-text-detector

Text Classification • 0.4B • Updated Jul 18 • 366
Team DACTYL at PAN 2026: Bayesian Data Mixing and Empirical X-risk Minimization for AI-text Detection

Paper • 2607.17382 • Published Jul 19 • 1

RIP Dataset

ShantanuT01/RIP-Dataset

Viewer • Updated Oct 23, 2024 • 19k • 25
Which LLMs are Difficult to Detect? A Detailed Analysis of Potential Factors Contributing to Difficulties in LLM Text Detection

Paper • 2410.14875 • Published Oct 18, 2024

DACTYL

Contains classifiers, fine-tuned SLMs, and datasets used for the Diverse Adversarial Corpus of Texts Yielded from Large language models.

DACTYL Classifiers

Collection

Trained AI-generated text classifiers. Pretrained means using binary cross entropy loss, finetuned refers to deep X-risk optimized classifiers. • 10 items • Updated Sep 24, 2025
DACTYL Finetuned SLMs

Collection

These are models that were continued pretrained on a domain-specific corpus. • 18 items • Updated Mar 2
ShantanuT01/DACTYL

Viewer • Updated Sep 24, 2025 • 655k • 73
ShantanuT01/DACTYL-Pretraining

Viewer • Updated Sep 25, 2025 • 416k • 47

DACTYL Classifiers

Trained AI-generated text classifiers. Pretrained means using binary cross entropy loss, finetuned refers to deep X-risk optimized classifiers.

ShantanuT01/dactyl-deberta-v3-large-pretrained

Text Classification • 0.4B • Updated Sep 24, 2025 • 4
ShantanuT01/dactyl-ai-text-detector

Text Classification • 0.4B • Updated Mar 4 • 1.11k
ShantanuT01/dactyl-modernbert-base-pretrained

Text Classification • 0.1B • Updated Sep 24, 2025 • 2
ShantanuT01/dactyl-modernbert-base-finetuned

Text Classification • 0.1B • Updated Sep 24, 2025 • 12

DACTYL Finetuned SLMs

These are models that were continued pretrained on a domain-specific corpus.

ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-training

Text Generation • 1B • Updated Apr 29, 2025 • 4
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-validation

Text Generation • 1B • Updated Apr 29, 2025 • 1
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-testing

Text Generation • 1B • Updated Apr 29, 2025 • 3
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-reviews-validation

Text Generation • 1B • Updated Apr 29, 2025 • 4

R255

R255 Project Models and Datasets

ShantanuT01/Llama-3.2-1B-Instruct-yelp-finetuned

1B • Updated Mar 27, 2025 • 5
ShantanuT01/Llama-3.2-1B-Instruct-abliterated-yelp-finetuned

Text Generation • 1B • Updated Mar 28, 2025 • 13 •
ShantanuT01/R255-Finetuning-Datasets

Preview • Updated Mar 28, 2025 • 23
ShantanuT01/Llama-3.2-1B-Instruct-abliterated-wp-finetuned

1B • Updated Mar 28, 2025 • 8