Unverified50% confidenceFactExact time
R2VC's generator undergoes supervised fine-tuning (SFT) and DPO alignment training to produce multiple diverse structured verdict candidates.
1
Sources
50%
Confidence
Long-term
Relevance
9/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedR2VC introduces an external NLI cross-encoder to filter among multiple candidates based on retrieved evidence.72% similarUnverified指令跟随能力依赖于监督微调（SFT）阶段和RLHF阶段的偏好对齐优化68% similarUnverifiedR2VC uses a lightweight sequence-level calibrator for confidence estimation and selective abstention.67% similarUnverifiedSupervised fine-tuning (SFT) trains a model using 'prompt–response' example pairs, transforming it from outputting gibberish to being capable of meaningful question answering.66% similarUnverifiedResearchers transformed the key features identified by SHAP into concept sets for TCAV (Testing with Concept Activation Vectors) to probe the multilingual XLM-R model and various language-specific encoders.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/919701API
curl https://kongchang.com/api/v1/knowledge/claims/919701MCP
get_claim(id=919701)