LLM popularity judgments align more closely with pretraining data exposure counts than with Wikipedia popularity, with stronger effects in pairwise comparisons and larger models.
Proceedings of the Annual ACM Symposium on Theory of Computing , year =
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
representative citing papers
CDVM formulates data pruning as maximizing total data influence while constraining excessive contributions to any single test point, yielding robust performance on the OpenDataVal benchmark in low-data regimes.
LLMs display high variance and major accuracy drops on GSM-Symbolic variants of grade-school math problems, indicating they replicate training patterns rather than execute logical reasoning.
citing papers explorer
-
Pretraining Exposure Explains Popularity Judgments in Large Language Models
LLM popularity judgments align more closely with pretraining data exposure counts than with Wikipedia popularity, with stronger effects in pairwise comparisons and larger models.
-
Constraint-Data-Value-Maximization: Utilizing Data Attribution for Effective Data Pruning in Low-Data Environments
CDVM formulates data pruning as maximizing total data influence while constraining excessive contributions to any single test point, yielding robust performance on the OpenDataVal benchmark in low-data regimes.
-
GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models
LLMs display high variance and major accuracy drops on GSM-Symbolic variants of grade-school math problems, indicating they replicate training patterns rather than execute logical reasoning.