Failed episodes exhibit systematically more negative temporal curvature kappa across all six task-LLM pairings, with correct minus incorrect mean differences of: Math (Llama-3.1-8B) +0.047, Code (Qwen3-14B) +0.148, Retail (Qwen3-14B) +0.030, Airline (Qwen3-14B) +0.049, Retail (Qwen3-32B) +0.035, and Airline (Qwen3-32B) +0.088.

来源：paper:Page 4, Section 4.2, Paragraph 'Failed episodes exhibit systematically more negative kappa'

报告指标与观测值

curvature_difference_correct_minus_incorrect

meas-diff-kappa-math-8b

论文报告 0.047 scalar

实际观测 — scalar

curvature_difference_correct_minus_incorrect

meas-diff-kappa-code-14b

论文报告 0.148 scalar

实际观测 — scalar

curvature_difference_correct_minus_incorrect

meas-diff-kappa-retail-14b

论文报告 0.03 scalar

实际观测 — scalar

curvature_difference_correct_minus_incorrect

meas-diff-kappa-airline-14b

论文报告 0.049 scalar

实际观测 — scalar

curvature_difference_correct_minus_incorrect

meas-diff-kappa-retail-32b

论文报告 0.035 scalar

实际观测 — scalar

curvature_difference_correct_minus_incorrect

meas-diff-kappa-airline-32b

论文报告 0.088 scalar

实际观测 — scalar

这条 Claim 暂无已发布的不可变 Assessment。

这条 Claim 暂无关联执行。