AIThe Decoderabout 1 hour ago
Study Reveals Nuance in Chinese AI Models
5 min read
Vals AI study finds Chinese models show political complexity, not just government alignment. Scores range from -1 to +1, with GLM-5.3 Flash confirming human rights violations. Thomson 1.0 Small fine-tuning drops pro-China bias from 84.6 to 15.0 percent. Western models occasionally adopt official Chinese positions.
Level
Hype check
Tap to vote and see what everyone thinks.
