Lukas Helff

LukasHug

22 17 52

https://www.ml.informatik.tu-darmstadt.de/people/lhelff/index.html

lukashelff

AI & ML interests

I am a PhD student in the AI and ML Lab at TU Darmstadt, specializing in deep learning and computer vision. My research primarily revolves around visual and logical reasoning using deep neural networks, symbolic AI, and Neural-Symbolic AI.

Recent Activity

liked a model 7 days ago

Soofi-Project/Soofi-S-Base

updated a dataset 21 days ago

AIML-TUDA/slr-leaderboard-requests

liked a model 29 days ago

Qwen/Qwen3.8-27B-FP8

View all activity

Organizations

liked a model 7 days ago

Soofi-Project/Soofi-S-Base

Text Generation • 32B • Updated Jul 18 • 512 • 103

liked a model 29 days ago

Qwen/Qwen3.8-27B-FP8

Image-Text-to-Text • 28B • Updated Aug 14 • 7.6M • 841

liked 2 models 2 months ago

Soofi-Project/Soofi-S-Isar-Preview

Text Generation • 32B • Updated Jul 17 • 1.32k • 16

Soofi-Project/Soofi-S-Rhine-Preview

Text Generation • 32B • Updated Jul 17 • 1.15k • 19

liked 2 models 3 months ago

AIML-TUDA/Olmo-3.1-7B-Think

Text Generation • 528k • Updated Jun 26 • 28 • 2

AIML-TUDA/OlmoLogic-7B-Think

Text Generation • 528k • Updated Jun 26 • 555 • 4

liked a dataset 3 months ago

AIML-TUDA/KletterMix-12B

Viewer • Updated Jun 4 • 11.2M • 3 • 3

liked a Space 4 months ago

SLR-Bench Leaderboard - Reward Hacking in Reasoning Models

🎯

Reward shortcut behavior in LLMs via IPT

liked a dataset 4 months ago

AIML-TUDA/SLR-Homes

Viewer • Updated May 13 • 500 • 24 • 1

liked a Space 4 months ago

Isomorphic Perturbation Testing

🔍

Evaluate rule hypotheses for genuine reasoning vs shortcuts

liked a dataset 10 months ago

AIML-TUDA/Rail2Country

Viewer • Updated May 21 • 4.8k • 30 • 3

liked a model 11 months ago

AIML-TUDA/Llama-3.1-8B-SLR

8B • Updated Oct 9, 2025 • 3 • 1

liked 2 datasets 11 months ago

AIML-TUDA/SLR-Bench-French

Viewer • Updated May 13 • 38.5k • 389 • 1

AIML-TUDA/SLR-Bench-Spanish

Viewer • Updated May 13 • 38.5k • 406 • 1

liked a dataset 12 months ago

AIML-TUDA/SLR-Bench-German

Viewer • Updated May 13 • 38.5k • 412 • 2

liked a Space about 1 year ago

VerifiableRewardsForScalableLogicalReasoning

🚀

Evaluate logical rules with a validation program

liked a dataset about 1 year ago

AIML-TUDA/SLR-Bench

Viewer • Updated May 22 • 38.5k • 1k • 5

liked 3 models over 1 year ago