Unverified50% confidenceBenchmarkExact time
LoRA performs comparably to or even better than full fine-tuning on RoBERTa, DeBERTa, GPT-2 and GPT-3, for example reaching 74.0% on WikiSQL and 91.7% on MultiNLI.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMMLU scores jumped from approximately 70% for GPT-3.5 to 86% for GPT-4, and then to above 90% for subsequent models.68% similarUnverifiedGPT-5.4 was reversed by a "trusted source" on 44.7% of questions, Grok-4.20 had a reversal rate as high as 87.5%, while Gemini-3.1-Pro was nearly unaffected at just 0.6%.67% similarUnverifiedWhen paired with GPT-4o, FinDialogLens achieves 92.1% accuracy on final price extraction.65% similarUnverifiedGPT-6 Astra scored a high 90% (72/80) on the KingBench 3 benchmark.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1047751API
curl https://kongchang.com/api/v1/knowledge/claims/1047751MCP
get_claim(id=1047751)