‘LLMs still find it difficult to pull off complex reasoning tasks’
BENGALURU: Cognizant AI CTO Babak Hodjat said at the Nasdaq-listed company’s AI Lab here that as enterprises race to embed large language models (LLMs) deeper into business operations, a key question remains unresolved: how much can these systems be trusted to reason correctly at scale? While today’s LLMs are powerful, he said, they tend to…