Ubiquitous Knowledge Processing Lab

university

https://www.informatik.tu-darmstadt.de/ukp

UKPLab

AI & ML interests

The Ubiquitous Knowledge Processing Lab researches natural language processing, text mining, eLearning, and digital humanities.

timbmg

authored a paper 3 months ago

The Good, the Bad and the Constructive: Automatically Measuring Peer Review's Utility for Authors

Paper • 2509.04484 • Published Aug 31 • 1

imbesat-rizvi

updated a model 4 months ago

UKPLab/Qwen2.5-3b-spare-prm-math

Text Generation • 3B • Updated Aug 21 • 8

imbesat-rizvi

published a model 4 months ago

UKPLab/Qwen2.5-3b-spare-prm-math

Text Generation • 3B • Updated Aug 21 • 8

haritzpuerto

authored 2 papers 5 months ago

Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers

Paper • 2506.15674 • Published Jun 18 • 2

C-SEO Bench: Does Conversational SEO Work?

Paper • 2506.11097 • Published Jun 6 • 2

imbesat-rizvi

updated a model 6 months ago

UKPLab/Llama-3-8b-spare-prm-math

Text Generation • 8B • Updated Jun 19 • 151

imbesat-rizvi

updated a dataset 6 months ago

UKPLab/sparp

Viewer • Updated Jun 5 • 417k • 117 • 1

imbesat-rizvi

published a model 6 months ago

UKPLab/Llama-3-8b-spare-prm-math

Text Generation • 8B • Updated Jun 19 • 151

iNeil77

authored a paper 6 months ago

Massively Multilingual Adaptation of Large Language Models Using Bilingual Translation Data

Paper • 2506.00469 • Published May 31 • 4

haritzpuerto

posted an update 6 months ago

Post

360

📜 Accepted at ACL 2025! Fine-Tuning on Diverse Reasoning Chains Drives Within-Inference CoT Refinement in LLMs
We propose to fine-tune LLMs to generate diverse chains of thought (DCoT) in a single inference step. This enables within-inference refinement of the cots, no external feedback needed!
🔗 https://arxiv.org/abs/2407.03181