Expired50% confidenceFactTime unknown
Deep thinking mode models are trained via reinforcement learning (such as RLHF or GRPO) to first output thinking tokens before the final answer.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
Alibaba Cloud Bailian Platform Hands-On: Complete Tutorial from API Calls to Multi-Turn Conversations
bilibiliAI大模型基地
Related Claims
UnverifiedAt the training level, reasoning-enhanced models typically leverage reinforcement learning algorithms such as GRPO (Group Relative Policy Optimization), using the correctness of the final answer as the reward signal72% similarUnverifiedModels such as o1 and DeepSeek-R1 internalize chain-of-thought as a training objective, generating a large amount of internal draft reasoning before producing the final answer.72% similarVerifiedThe o-series models combine chain-of-thought reasoning with reinforcement learning, with the model generating internal thinking steps before producing an answer.69% similarUnverifiedReinforcement Learning from Verifiable Rewards (RLVR) has become an important path for enhancing model reasoning capabilities and is one of the key factors behind the capability leaps of several frontier reasoning models68% similarUnverifiedModern reasoning models generally use RLHF or its variant RLAIF for alignment training, where the model learns during training when to stop reasoning and when to output an answer.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/54873API
curl https://kongchang.com/api/v1/knowledge/claims/54873MCP
get_claim(id=54873)