Unverified50% confidenceFactExact time
Prompt Tuning was proposed by Lester et al. in 2021, prepending a set of learnable soft prompt vectors before the input embeddings.
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPrompt engineering emerged in the early period of large language model popularization in 2022–2023, with core methods including Zero-shot prompting, Few-shot prompting, Chain-of-Thought, and other techniques.64% similarUnverifiedMidjourney uses CLIP's text encoder to convert prompts into conditioning vectors that guide the denoising process of the diffusion model.63% similarUnverifiedFastEmbed支持sentence-transformers/all-MiniLM-L6-v2和mixedbread-ai/mxbai-embed-large-v1等多种预训练模型62% similarUnverifiedThe concept of Vibe Coding was originally proposed by Andrej Karpathy between 2024 and 2025, referring to a more intuitive and lower-friction way of coding.62% similarUnverifiedThe emergence of BERT in 2018 transformed the resume matching field, as Transformer-based pre-trained models can encode text into high-dimensional semantic vectors to identify near-synonymous expressions.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/888477API
curl https://kongchang.com/api/v1/knowledge/claims/888477MCP
get_claim(id=888477)