Teaching language models to reason algorithmically

Large language models (LLMs), such as GPT-3 and PaLM, have shown impressive progress in recent years, which have been driven by scaling up models and training data sizes. Nonetheless, a long standing debate has been whether LLMs can reason symbolically (i.e., manipulating symbols based on logical rules). For example, LLMs are able to perform simple arithmetic operations when numbers are small, but struggle to perform …

data google gpt gpt-3 graduate language language models large language large language models llms palm progress reason research research scientist scaling scaling up teaching training training data