LoRA fine-tuning on control systems course data increases lexical and semantic similarity between model responses and course reference answers at both 3B and 7B model sizes¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q2
Currency · newest study 2026 · checked against its sources 2026-10-20 (automated, 2 of 2 entries) · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? LoRA-fine-tuned Qwen2.5 models outperformed their corresponding base models on most ROUGE and BERTScore metrics on the 54-item test set. → Shaowen Lu 2026 q2 i? Paired bootstrap 95% confidence intervals for ROUGE improvements lay entirely above zero at most settings, indicating stable gains. → Shaowen Lu 2026 (2)

Evidence¶

Shaowen Lu 2026¶

Shaowen Lu, Chengxu Liu, Ping Zhou and Tao Yang. (2026). LoRA Fine-Tuned Models for Control Systems Course Q&A: A Multidimensional Evaluation of Model Scale and Rank Effects. https://arxiv.org/abs/2609.13918 doi:10.48550/arxiv.2609.13918

q2 · i? · causal · r2

Test-set comparison of base and LoRA-fine-tuned Qwen2.5-3B and 7B models on 54 unseen course questions under identical prompts and deterministic decoding, reported via ROUGE-1/2/L and BERTScore-F1. The article reports that "The LoRA-fine-tuned models outperformed their corresponding base models on most metrics."

"The LoRA-fine-tuned models outperformed their corresponding base models on most metrics. Fine-tuning on course data therefore increased both lexical overlap and semantic proximity to the reference answers."

Shaowen Lu 2026 (2)¶

Shaowen Lu, Chengxu Liu, Ping Zhou and Tao Yang. (2026). LoRA Fine-Tuned Models for Control Systems Course Q&A: A Multidimensional Evaluation of Model Scale and Rank Effects. https://arxiv.org/abs/2609.13918

q2 · i? · causal · r2

Paired bootstrap resampling (B=10,000 iterations over the 54-item test set) of ROUGE gains of fine-tuned over base models. The article reports that for the 3B model "the confidence intervals for ROUGE-1, ROUGE-2, and ROUGE-L were above zero" at all three ranks.

"For the 3B model, the confidence intervals for ROUGE-1, ROUGE-2, and ROUGE-L were above zero at r= 4 , r= 8 , and r= 16 ."

Discussion¶

Related Claims¶

• Bootstrap stability of LoRA gains is not uniform: the ROUGE-1 confidence interval crossed zero at 7B-r4, so that increase is a trend rather than a stable improvement — related
• Base-model size effects persist after LoRA adaptation: the 7B model gained more on ROUGE-2 and ROUGE-L, yet 3B-r16 outperformed 7B-r4 on ROUGE-L and BERTScore-F1 — related
• LoRA fine-tuning makes the Solution-Method-Teaching Points instructional format more stable, with Solution and Method coverage reaching 1.00 at r=8 and r=16 for the 3B model — related
• Increasing LoRA rank raises absolute ROUGE-L gains but produces diminishing returns in performance per million adapter parameters — related
• Fine-tuning Llama2-7B with parameter-efficient methods yields unsatisfactory results for measuring subject-matter teaching practices, only marginally improving the majority baseline — related
• Adaptation gains grow monotonically with model capacity under the same LoRA recipe, though three data points cannot establish a scaling law — related
• LoRA adaptation significantly increases closed-book accuracy on the PhysicsMate test set at all three Qwen3 scales — related
• Cluster membership showed a strong multivariate effect on standard text-similarity metrics (Pillai's trace = 0.85, F(12, 248) = 15.31, p < 0.001) — related