Contextual noise from irrelevant textbooks broadly degrades KC planning, most in Validity and Efficiency¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Adding irrelevant textbook titles costs Qwen3-30B-A3B (CoT) 54.9 percentage points in Validity and 35.7 points in Efficiency in Basic Education. → Yu Liu 2026

Evidence¶

Yu Liu 2026¶

Yu Liu, Zeming Liu, Tianle Zhang, Zihao Cheng, Yuhang Guo, Kehai Chen, Min Zhang, Yunhong Wang, Haifeng Wang. (2026). PersonaPath: Towards Knowledge-Centric Personalized Learning Path Planning. arXiv. https://arxiv.org/abs/2609.18861 doi:10.48550/arxiv.2609.18861

q2 · i? · design · r2

Controlled noise-injection experiment (Table 7) in which irrelevant textbook titles were added at varying proportions to the candidate pool for Basic Education personas. The article reports the printed percentage-point losses; no effect size is printed.

"Contextual noise degrades performance broadly, most notably in Validity and Efficiency. Qwen3-30B-A3B (CoT) loses 54.9 percentage points in Validity and 35.7 points in Efficiency."

Discussion¶

Related Claims¶

• Removing explicit mastery information from learner profiles mainly degrades Adaptivity — related
• Retrieval augmentation raises path Validity but not the joint Final Pass rate — related
• Chain-of-Thought prompting improves KC planning for Qwen models but its effect varies by model and educational stage — related
• Preserving book-level document structure during mid-training improves downstream performance beyond identical content split into sections (+1.02 mean) — related