The widespread adoption of large language models (LLMs) makes it important to
recognize their strengths and limitations. We argue that in order to develop a
holistic understanding of these systems we need to consider the problem that
they were trained to solve: next-word prediction over Internet text. By
recognizing the pressures that this task exerts we can make predictions about
the strategies that LLMs will adopt, allowing us to reason about when they will
succeed or fail. This approach - which we call the teleological approach -
leads us to identify three factors that we hypothesize will influence LLM
accuracy: the probability of the task to be performed, the probability of the
target output, and the probability of the provided input. We predict that LLMs
will achieve higher accuracy when these probabilities are high than when they
are low - even in deterministic settings where probability should not matter.
To test our predictions, we evaluate two LLMs (GPT-3.5 and GPT-4) on eleven
tasks, and we find robust evidence that LLMs are influenced by probability in
the ways that we have hypothesized. In many cases, the experiments reveal
surprising failure modes. For instance, GPT-4's accuracy at decoding a simple
cipher is 51% when the output is a high-probability word sequence but only 13%
when it is low-probability. These results show that AI practitioners should be
careful about using LLMs in low-probability situations. More broadly, we
conclude that we should not evaluate LLMs as if they are humans but should
instead treat them as a distinct type of system - one that has been shaped by
its own particular set of pressures.

大语言模型的广泛应用使得识别它们的优势和局限性变得重要。我们主张，为了全面理解这些系统，我们需要考虑它们在训练中解决的问题：互联网文本的下一个词预测。通过认识到这个任务所带来的压力，我们可以对大语言模型采用的策略进行预测，从而推断它们的成功或失败。这种方法，我们称之为目的论方法，使我们能够确定三个因素，我们假设这些因素会影响大语言模型的准确性：执行任务的概率、目标输出的概率和提供的输入的概率。我们预测，当这些概率较高时，大语言模型的准确性会更高，而当概率较低时，即使在确定性环境下，概率也不应该起作用。为了验证我们的预测，我们对两个大语言模型（GPT-3.5 和 GPT-4）进行了十一项任务的评估，我们找到了强有力的证据表明大语言模型受到我们假设的概率影响的方式。在许多情况下，实验揭示了令人惊讶的失效模式。例如，当输出是高概率单词序列时，GPT-4 解码简单密码的准确率为 51%，但当输出是低概率时，准确率为 13%。这些结果表明 AI 从业者在低概率情况下使用大语言模型时需要谨慎。更广泛地说，我们得出结论，我们不应该把大语言模型评估为人类，而应该把它们看作一类独特的系统 —— 这类系统经过了自己特定的压力塑造。