Unverified50% confidenceFactExact time
Another implementation of zero-shot classification is to directly call instruction-following large language models (such as GPT-4 and Claude), using prompts to describe classification rules and output category labels.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/5/2026
First Seen
Valid until: 1/3/2027
Sources
Related Entities
Related Claims
UnverifiedZero-shot classifiers (such as solutions based on CLIP or NLI) allow users to describe categories in natural language, offering greater flexibility but typically requiring a larger base model.71% similarUnverifiedZero-shot models require no labeled data; simply writing out labels in natural language is enough to produce classification results.70% similarVerifiedZero-shot classification relies on the semantic understanding capabilities of pretrained large language models such as NLI variants of BART and RoBERTa.69% similarUnverified规则层通过关键词、正则表达式、状态机处理表意固定无歧义的高频指令，可实现毫秒级路由且零算力消耗68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/976265API
curl https://kongchang.com/api/v1/knowledge/claims/976265MCP
get_claim(id=976265)