Learning New Facts with QLoRA: An Acquisition-Retention Frontier
Parameter-efficient fine-tuning is often assumed to preserve pretrained capabilities because it updates only a small number of parameters. We show that this assumption depends strongly on adapter capacity. We study factual acquisition in a controlled OpenStreetMap-derived benchmark where Qwen3-4B must acquire anonymized geographic associations while retaining unrelated capabilities. Comparing full fine-tuning (FFT) with quantized low-rank adaptation (QLoRA) at ranks 8, 16, 32, and 64, we find that rank induces a clear acquisition--retention frontier. Low-rank QLoRA preserves out-of-domain (OOD
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 47%Fine-tuning →
- FuzzySimilar title/name (fuzzy) · 59%aymericdamien/TopDeepLearning →
“Fuzzy title match (0.73): “Learning New Facts with QLoRA: An Acquisition-Retention Fron” ≈ “aymericdamien/TopDeepLearning””
- LinkedLinked via arxiv author · 85%Estelle Zheng →
“Learning New Facts with QLoRA: An Acquisition-Retention Frontier”
- LinkedLinked via arxiv author · 85%Sébastien Warichet →
“Learning New Facts with QLoRA: An Acquisition-Retention Frontier”
- LinkedLinked via arxiv author · 85%Emmanuel Helbert →
“Learning New Facts with QLoRA: An Acquisition-Retention Frontier”
- LinkedLinked via arxiv author · 85%Christophe Cerisara →
“Learning New Facts with QLoRA: An Acquisition-Retention Frontier”
- FuzzySimilar title/name (fuzzy) · 84%amitness/learning →
“Fuzzy title match (0.92): “Learning New Facts with QLoRA: An Acquisition-Retention Fron” ≈ “amitness/learning””