Unverified50% confidenceBenchmarkExact time
Performing task-specific reinforcement learning on a small 4B-parameter model improved its mapping reasoning accuracy by 15.7 percentage points.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFew-shot learning and Chain-of-Thought prompting techniques can significantly improve model performance on reasoning tasks.74% 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73% similarUnverified链式推理通过引导模型在给出最终答案前显式输出中间推理步骤，某些任务上准确率提升可达20-40个百分点73% similarUnverified结构化的任务描述（包含背景、目标、约束和期望输出格式四个维度）相比自然语言描述，能将模型的首次响应准确率提升 30%~50%71% similarUnverified在MATH数学推理数据集上，采用元提示策略的模型表现比标准提示方法提升了约17%71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/902748API
curl https://kongchang.com/api/v1/knowledge/claims/902748MCP
get_claim(id=902748)