Tag: LLM post-training techniques

AI Daily – 2025-05-01(Morning)

AI agent protocol DeepSeek-Prover-V2 DeepSeek-Prover-V2 theorem proving performance DiffTransformer long-text modeling GPT-4o excessive sycophancy behavior fix GPT-4o update rollback LLM post-training techniques Microsoft Phi-4-reasoning-plus model MoE inference optimization Phi-4 reasoning model Qwen3 by Tongyi Qianwen Qwen3-235B multilingual support