Unverified50% confidenceFactExact time
The success of GPT-3 validated the scalability and flexibility of the decoder-only architecture at ultra-large scale, after which mainstream research converged toward decoder-only.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Partially VerifiedThe GPT series adopts a Decoder-only architecture.79% similarUnverified代码推理能力的质变在 GPT-3 到 GPT-4 的代际跨越中被系统性观察到，符合规模化定律的涌现现象72% similarUnverifiedModern large language models (such as the GPT series) almost universally adopt a decoder-only architecture, using masked self-attention to ensure that autoregressive generation does not leak future token information.72% similarUnverifiedGPT-2采用纯解码器的Transformer堆叠结构，去掉了编码器-解码器的交叉注意力部分，专注于单向自回归语言建模69% similarUnverifiedMainstream generative large models such as GPT-4, LLaMA, and Qwen adopt a Decoder-only architecture; BERT and RoBERTa adopt an Encoder-only architecture; T5 and BART adopt an Encoder-Decoder architecture.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/916431API
curl https://kongchang.com/api/v1/knowledge/claims/916431MCP
get_claim(id=916431)