PiKa provides 30k-example datasets that allow base LLMs to surpass officially aligned versions on standard benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
PIKA: Expert-Level Synthetic Datasets for Post-Training Alignment from Scratch
PiKa provides 30k-example datasets that allow base LLMs to surpass officially aligned versions on standard benchmarks.