A research team, in collaboration with academics from ETH Zurich, Tel Aviv University and University of Haifa, created a benchmark called CryptanalysisBench to examine the cryptanalytic capabilities of large language models (LLMs). The system aims to assess model vulnerabilities and identify security risks, supporting the development of future defensive strategies.
CryptanalysisBench: a new benchmark for testing the cryptanalytic capabilities of large language models
A research team, in collaboration with academics from ETH Zurich, Tel Aviv University and University of Haifa, created a benchmark called CryptanalysisBench to examine the cryptanalytic capabilities of large language models (LLMs).
