Prati
Jannik Kossen
Naslov
Citirano
Citirano
Godina
Detecting hallucinations in large language models using semantic entropy
S Farquhar*, J Kossen*, L Kuhn*, Y Gal
Nature 630 (8017), 625-630, 2024
1872024
Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning
J Kossen, N Band, C Lyle, AN Gomez, T Rainforth, Y Gal
NeurIPS, 2021
1492021
Global green hydrogen-based steel opportunities surrounding high quality renewable energy and iron ore deposits
A Devlin, J Kossen, H Goldie-Jones, A Yang
Nature Communications 14 (1), 2578, 2023
972023
Structured Object-Aware Physics Prediction for Video Modeling and Planning
J Kossen, K Stelzner, M Hussing, C Voelcker, K Kersting
ICLR, 2020
722020
Active Testing: Sample-Efficient Model Evaluation
J Kossen, S Farquhar, Y Gal, T Rainforth
ICML, 2021
672021
In-context learning learns label relationships but is not conventional learning
J Kossen, T Rainforth, Y Gal
ICLR, 2024
54*2024
Wie Maschinen lernen
K Kersting, C Lampert, C Rothkopf
Wiesbaden: Springer, 2019
442019
Active Surrogate Estimators: An Active Learning Approach to Label-Efficient Model Evaluation
J Kossen, S Farquhar, Y Gal, T Rainforth
NeurIPS, 2022
202022
Semantic entropy probes: Robust and cheap hallucination detection in llms
J Kossen, J Han, M Razzak, L Schut, S Malik, Y Gal
arXiv preprint arXiv:2406.15927, 2024
132024
Kernel language entropy: Fine-grained uncertainty quantification for LLMs from semantic similarities
A Nikitin, J Kossen, Y Gal, P Marttinen
Advances in Neural Information Processing Systems 37, 8901-8929, 2025
92025
Three Towers: Flexible Contrastive Learning with Pretrained Image Models
J Kossen, M Collier, B Mustafa, X Wang, X Zhai, L Beyer, A Steiner, ...
NeurIPS, 2023
92023
Active Acquisition for Multimodal Temporal Data: A Challenging Decision-Making Task
J Kossen, C Cangea, E Vértes, A Jaegle, V Patraucean, I Ktena, ...
TMLR, 2023
72023
Semantic entropy probes: Robust and cheap hallucination detection in llms, 2024
J Kossen, J Han, M Razzak, L Schut, S Malik, Y Gal
URL https://arxiv. org/abs/2406.15927, 0
5
Estimating the hallucination rate of generative ai
A Jesson, N Beltran Velez, Q Chu, S Karlekar, J Kossen, Y Gal, ...
Advances in Neural Information Processing Systems 37, 31154-31201, 2025
42025
Klassifikation: Schubladendenken!
J Aberham, J Kossen
Wie Maschinen lernen: Künstliche Intelligenz verständlich erklärt, 45-52, 2019
32019
Marginal and joint cross-entropies & predictives for online Bayesian inference, active learning, and active sampling
A Kirsch, J Kossen, Y Gal
arXiv preprint arXiv:2205.08766, 2022
22022
Faltungsnetze: Neuronales Origami
J Kossen, ME Müller
Wie Maschinen lernen: Künstliche Intelligenz verständlich erklärt, 163-169, 2019
22019
Regression: Voll im Trend
J Kossen, ME Müller
Wie Maschinen lernen: Künstliche Intelligenz verständlich erklärt, 39-43, 2019
22019
Rethinking Aleatoric and Epistemic Uncertainty
FB Smith, J Kossen, E Trollope, M van der Wilk, A Foster, T Rainforth
arXiv preprint arXiv:2412.20892, 2024
12024
Fine-Tuning Large Language Models to Appropriately Abstain with Semantic Entropy
BA Tjandra, M Razzak, J Kossen, K Handa, Y Gal
arXiv preprint arXiv:2410.17234, 2024
12024
Sustav trenutno ne može provesti ovu radnju. Pokušajte ponovo kasnije.
Članci 1–20