Unverified50% confidenceFactExact time
The study used the Qwen2.5-Coder-7B-Instruct model, supplemented by the 1.5B model for sensitivity analysis, completing a total of 13,638 model requests.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe research team conducted fine-tuning experiments on two models, Qwen3.5-9B and Gemma4-12B-it.70% similarUnverifiedOn Gemma-2-9B and Qwen3-14B, the researchers successfully switched the facts read by the model by modifying early internal activation vectors while keeping the written instructions unchanged.66% similarUnverified思维链提示与少样本示例可提升7B至14B参数量级模型在复杂编码任务上的成功率66% similarUnverifiedA study based on 1.2 million model calls systematically quantified the inter-run variance of semantic query engines65% similarUnverifiedThe research team fine-tuned the model on 8.62K synthetic research dialogues.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1106287API
curl https://kongchang.com/api/v1/knowledge/claims/1106287MCP
get_claim(id=1106287)