LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
Minimum Error Rate Training in Statistical Machine Translation
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
POTracker fine-tunes Qwen2.5-7B-Instruct with POTrackerLoss to generate standard-compliant power outage reports, achieving up to 51% accuracy improvement and 86.47% structural accuracy on 1,000 reports.
citing papers explorer
-
LIMO: Less is More for Reasoning
LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
-
POTracker: Optimizing Large Language Models for Standard-Compliant Power Outage Report Generation
POTracker fine-tunes Qwen2.5-7B-Instruct with POTrackerLoss to generate standard-compliant power outage reports, achieving up to 51% accuracy improvement and 86.47% structural accuracy on 1,000 reports.