Emergent Inabilities? Inverse Scaling Over the Course of Pretraining
Submission Type: Regular Short Paper
Submission Track: Language Modeling and Analysis of Language Models
Submission Track 2: Theme Track: Large Language Models and the Future of NLP
Keywords: language models, inverse scaling, transformers, training dynamics
TL;DR: Language models can show inverse scaling not only as a function of number of model parameters, but also of model training data.
Abstract: Does inverse scaling only occur as a function of model size, or can it also occur over the course of training? We carry out an exploratory study investigating whether the performance of language models on specific tasks can decrease (while general performance remains high) during training on the language modeling task. We find 8 tasks on which Pythia 12B (Biderman et al., 2023) shows decreased performance over the course of training. Five of these tasks (TruthfulQA-MC1, TruthfulQA-MC2, Hindsight Neglect, Memo Trap, and Pattern Match Suppression) additionally show a consistent relationship whereby larger language models show a greater decrease in performance the more they are trained, despite showing standard (positive) scaling overall. This highlights the importance of testing performance at all relevant benchmarks any time models are trained on additional data, even if their overall performance improves.
Submission Number: 3937
Loading