Fine-tuning reasoning models on answer-only data induces reasoning-trace collapse where valid traces disappear while answer performance stays high, and simple loss-masking can mitigate it.
A Comprehensive Survey of Small Language Models in the Era of Large Language Models: Techniques, Enhancements, Applications, Collaboration with LLMs, and Trustworthiness.ACM Trans
3 Pith papers cite this work, alongside 52 external citations. Polarity classification is still indexing.
3
Pith papers citing it
52
external citations · OpenAlex
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Checkpoint monitoring during sequential LoRA adaptation of SLMs reveals instability patterns via reference set diagnostics that standard task metrics can miss.
Small language models can run RAG generation on-device without GPUs in reasonable time.
citing papers explorer
-
Reasoning-Trace Collapse: Evaluating the Loss of Explicit Reasoning During Fine-Tuning
Fine-tuning reasoning models on answer-only data induces reasoning-trace collapse where valid traces disappear while answer performance stays high, and simple loss-masking can mitigate it.
-
Continual Learning for Sequential Personalization of Small Language Models: A Stability Monitoring Analysis
Checkpoint monitoring during sequential LoRA adaptation of SLMs reveals instability patterns via reference set diagnostics that standard task metrics can miss.
-
Little Brains, Big Feats: Exploring Compact Language Models
Small language models can run RAG generation on-device without GPUs in reasonable time.