# Measuring Systematic Generalization in Neural Proof Generation with Transformers

arXiv: https://arxiv.org/abs/2009.14786 · HTML: https://mlsanity.com/paper/2009.14786/

## Key facts

- Cited by 43 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Sep 2020.

## Built on by (most cited first, 30 of 43)

- [Solving math word problems with process- and outcome-based feedback](https://mlsanity.com/paper/2211.14275/) (2022)
- [The Impact of Positional Encoding on Length Generalization in Transformers](https://mlsanity.com/paper/2305.19466/) (2023)
- [A Review on Language Models as Knowledge Bases](https://mlsanity.com/paper/2204.06031/) (2022)
- [State-of-the-art generalisation research in NLP: A taxonomy and review](https://mlsanity.com/paper/2210.03050/) (2022)
- [On the Paradox of Learning to Reason from Data](https://mlsanity.com/paper/2205.11502/) (2022)
- [Testing the General Deductive Reasoning Capacity of Large Language Models Using OOD Examples](https://mlsanity.com/paper/2305.15269/) (2023)
- [LAMBADA: Backward Chaining for Automated Reasoning in Natural Language](https://mlsanity.com/paper/2212.13894/) (2022)
- [Generating Natural Language Proofs with Verifier-Guided Search](https://mlsanity.com/paper/2205.12443/) (2022)
- [Systematic Generalization with Edge Transformers](https://mlsanity.com/paper/2112.00578/) (2021)
- [The Neural Race Reduction: Dynamics of Abstraction in Gated Networks](https://mlsanity.com/paper/2207.10430/) (2022)
- [Enhancing Reasoning Capabilities of LLMs via Principled Synthetic Logic Corpus](https://mlsanity.com/paper/2411.12498/) (2024)
- [RobustLR: Evaluating Robustness to Logical Perturbation in Deductive Reasoning](https://mlsanity.com/paper/2205.12598/) (2022)
- [Position: Understanding LLMs Requires More Than Statistical Generalization](https://mlsanity.com/paper/2405.01964/) (2024)
- [Pushing the Limits of Rule Reasoning in Transformers through Natural Language Satisfiability](https://mlsanity.com/paper/2112.09054/) (2021)
- [RECKONING: Reasoning through Dynamic Knowledge Encoding](https://mlsanity.com/paper/2305.06349/) (2023)
- [Learning Deductive Reasoning from Synthetic Corpus based on Formal Logic](https://mlsanity.com/paper/2308.07336/) (2023)
- [AbductionRules: Training Transformers to Explain Unexpected Inputs](https://mlsanity.com/paper/2203.12186/) (2022)
- [Thinking Aloud: Dynamic Context Generation Improves Zero-Shot Reasoning Performance of GPT-2](https://mlsanity.com/paper/2103.13033/) (2021)
- [multiPRover: Generating Multiple Proofs for Improved Interpretability in Rule Reasoning](https://mlsanity.com/paper/2106.01354/) (2021)
- [A Survey on Compositional Learning of AI Models: Theoretical and Experimental Practices](https://mlsanity.com/paper/2406.08787/) (2024)
- [NOAHQA: Numerical Reasoning with Interpretable Graph Question Answering Dataset](https://mlsanity.com/paper/2109.10604/) (2021)
- [Summarization Programs: Interpretable Abstractive Summarization with Neural Modular Trees](https://mlsanity.com/paper/2209.10492/) (2022)
- [Learning Symbolic Rules for Reasoning in Quasi-Natural Language](https://mlsanity.com/paper/2111.12038/) (2021)
- [A Survey on Symbolic Knowledge Distillation of Large Language Models](https://mlsanity.com/paper/2408.10210/) (2024)
- [Does Pre-training Induce Systematic Inference? How Masked Language Models Acquire Commonsense Knowledge](https://mlsanity.com/paper/2112.08583/) (2021)
- [Diff-Explainer: Differentiable Convex Optimization for Explainable Multi-hop Inference](https://mlsanity.com/paper/2105.03417/) (2021)
- [DeepA2: A Modular Framework for Deep Argument Analysis with Pretrained Neural Text2Text Language Models](https://mlsanity.com/paper/2110.01509/) (2021)
- [Dyna-bAbI: unlocking bAbI's potential with dynamic synthetic benchmarking](https://mlsanity.com/paper/2112.00086/) (2021)
- [Random Scaling of Emergent Capabilities](https://mlsanity.com/paper/2502.17356/) (2025)
- [Breakpoint Transformers for Modeling and Tracking Intermediate Beliefs](https://mlsanity.com/paper/2211.07950/) (2022)

## Abstract (arXiv)

We are interested in understanding how well Transformer language models (TLMs) can perform reasoning tasks when trained on knowledge encoded in the form of natural language. We investigate their systematic generalization abilities on a logical reasoning task in natural language, which involves reasoning over relationships between entities grounded in first-order logical proofs. Specifically, we perform soft theorem-proving by leveraging TLMs to generate natural language proofs. We test the generated proofs for logical consistency, along with the accuracy of the final inference. We observe length-generalization issues when evaluated on longer-than-trained sequences. However, we observe TLMs improve their generalization performance after being exposed to longer, exhaustive proofs. In addition, we discover that TLMs are able to generalize better using backward-chaining proofs compared to their forward-chaining counterparts, while they find it easier to generate forward chaining proofs. We observe that models that are not trained to generate proofs are better at generalizing to problems based on longer proofs. This suggests that Transformers have efficient internal reasoning strategies that are harder to interpret. These results highlight the systematic generalization behavior of TLMs in the context of logical reasoning, and we believe this work motivates deeper inspection of their underlying reasoning strategies.