Self-training Language Models for Arithmetic Reasoning
- URL: http://arxiv.org/abs/2407.08400v3
- Date: Wed, 23 Oct 2024 20:43:02 GMT
- Title: Self-training Language Models for Arithmetic Reasoning
- Authors: Marek Kadlčík, Michal Štefánik,
- Abstract summary: We explore the potential of improving models' reasoning capabilities without new data.
We find that models can substantially improve in both single-round (offline) and online self-training.
- Score: 0.0
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Recent language models achieve impressive results in tasks involving complex multistep reasoning, but scaling these capabilities further traditionally requires expensive collection of more annotated data. In this work, we explore the potential of improving models' reasoning capabilities without new data, merely using automated feedback to the validity of their predictions in arithmetic reasoning (self-training). In systematic experimentation across six different arithmetic reasoning datasets, we find that models can substantially improve in both single-round (offline) and online self-training, reaching a correct result in +13.9% and +25.9% more cases, respectively, underlining the importance of actuality of self-training feedback. We further find that in the single-round, offline self-training, traditional supervised training can deliver gains comparable to preference optimization, but in online self-training, preference optimization methods largely outperform supervised training thanks to their superior stability and robustness on unseen types of problems.
Related papers
- Multiagent Finetuning: Self Improvement with Diverse Reasoning Chains [114.76612918465948]
Large language models (LLMs) have achieved remarkable performance in recent years but are fundamentally limited by the underlying training data.
We propose a complementary approach towards self-improvement where finetuning is applied to a multiagent society of language models.
arXiv Detail & Related papers (2025-01-10T04:35:46Z) - Self-Improvement in Language Models: The Sharpening Mechanism [70.9248553790022]
We offer a new perspective on the capabilities of self-improvement through a lens we refer to as sharpening.
Motivated by the observation that language models are often better at verifying response quality than they are at generating correct responses, we formalize self-improvement as using the model itself as a verifier during post-training.
We analyze two natural families of self-improvement algorithms based on SFT and RLHF.
arXiv Detail & Related papers (2024-12-02T20:24:17Z) - Enhancing LLM Reasoning via Critique Models with Test-Time and Training-Time Supervision [120.40788744292739]
We propose a two-player paradigm that separates the roles of reasoning and critique models.
We first propose AutoMathCritique, an automated and scalable framework for collecting critique data.
We demonstrate that the critique models consistently improve the actor's performance on difficult queries at test-time.
arXiv Detail & Related papers (2024-11-25T17:11:54Z) - What Do Learning Dynamics Reveal About Generalization in LLM Reasoning? [83.83230167222852]
We find that a model's generalization behavior can be effectively characterized by a training metric we call pre-memorization train accuracy.
By connecting a model's learning behavior to its generalization, pre-memorization train accuracy can guide targeted improvements to training strategies.
arXiv Detail & Related papers (2024-11-12T09:52:40Z) - Self-Training with Direct Preference Optimization Improves Chain-of-Thought Reasoning [5.487210426671288]
In this work, we demonstrate that the reasoning abilities of small-scale LMs can be enhanced through self-training.
We also show that the conventional self-training can be further augmented by a preference learning algorithm called Direct Preference Optimization.
arXiv Detail & Related papers (2024-07-25T17:59:16Z) - Towards Stable Machine Learning Model Retraining via Slowly Varying Sequences [6.067007470552307]
We propose a model-agnostic framework for finding sequences of models that are stable across retraining iterations.
We develop a mixed-integer optimization formulation that is guaranteed to recover optimal models.
We find that, on average, a 2% reduction in predictive power leads to a 30% improvement in stability.
arXiv Detail & Related papers (2024-03-28T22:45:38Z) - Fantastic Gains and Where to Find Them: On the Existence and Prospect of
General Knowledge Transfer between Any Pretrained Model [74.62272538148245]
We show that for arbitrary pairings of pretrained models, one model extracts significant data context unavailable in the other.
We investigate if it is possible to transfer such "complementary" knowledge from one model to another without performance degradation.
arXiv Detail & Related papers (2023-10-26T17:59:46Z) - Entailment as Robust Self-Learner [14.86757876218415]
We design a prompting strategy that formulates a number of different NLU tasks as contextual entailment.
We propose the Simple Pseudo-Label Editing (SimPLE) algorithm for better pseudo-labeling quality in self-training.
arXiv Detail & Related papers (2023-05-26T18:41:23Z) - INGENIOUS: Using Informative Data Subsets for Efficient Pre-Training of
Language Models [40.54353850357839]
We show how we can employ submodular optimization to select highly representative subsets of the training corpora.
We show that the resulting models achieve up to $sim99%$ of the performance of the fully-trained models.
arXiv Detail & Related papers (2023-05-11T09:24:41Z) - Self-Distillation for Further Pre-training of Transformers [83.84227016847096]
We propose self-distillation as a regularization for a further pre-training stage.
We empirically validate the efficacy of self-distillation on a variety of benchmark datasets for image and text classification tasks.
arXiv Detail & Related papers (2022-09-30T02:25:12Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.