Follow
Kazuki Egashira
Kazuki Egashira
Verified email at hal.t.u-tokyo.ac.jp
Title
Cited by
Cited by
Year
Exploiting LLM Quantization
K Egashira, M Vero, R Staab, J He, M Vechev
NeurIPS 2024 (Conference on Neural Information Processing Systems), 2024
142024
JMMMU: A Japanese Massive Multi-discipline Multimodal Understanding Benchmark for Culture-aware Evaluation
S Onohara*, A Miyai*, Y Imajuku*, K Egashira*, J Baek*, X Yue, G Neubig, ...
NAACL 2025 (Conference of the Nations of the Americas Chapter of the …, 2024
2024
Negative Learning to Prevent Undesirable Misclassification
K Egashira, A Miyai, Q Yu, G Irie, K Aizawa
IEICE TRANSACTIONS on Information and Systems 107 (1), 144-147, 2024
2024
Mind the Gap: A Practical Attack on GGUF Quantization
K Egashira, R Staab, M Vero, J He, M Vechev
ICLR 2025 Workshop on Building Trust in Language Models and Applications, 0
The system can't perform the operation now. Try again later.
Articles 1–4