GLM-5.3 raises researcher fears the Chinese model can fake AI safety testsscmp.com2 points·theanonymousone··0 commentsOpen articleSaveView on HN