Is More Data Worth the Cost? Dataset Scaling Laws in a Tiny Attention-Only Decoder
This provides practical guidance for balancing dataset size and cost in compute-restricted environments like small labs, though it is incremental as it applies known scaling laws to a controlled setting.
The study investigated the effect of dataset size on performance in a small-scale attention-only decoder, finding that using only about 30% of the training data achieves approximately 90% of the full-data validation accuracy, demonstrating diminishing returns.
Training Transformer language models is expensive, as performance typically improves with increasing dataset size and computational budget. Although scaling laws describe this trend at large scale, their implications in controlled, smaller-scale settings remain less explored. In this work, we isolate dataset-size effects using a strongly reduced attention-only decoder architecture. By training on progressively larger power-of-two subsets, we observe smooth performance improvements accompanied by clear diminishing returns, consistent with scaling-law behavior. Using only about 30% of the training data is sufficient to reach approximately 90% of the full-data validation token-level accuracy. These results provide actionable insights into dataset scaling in a controlled, component-isolated setting and offer practical guidance for balancing dataset size and computational cost in compute- and data-restricted environments, such as small research labs and exploratory model development.