Introduction

This chapter covers how to make completions more reliable, as well as how to implement checks to ensure that outputs are reliable.

To a certain extent, most of the previous techniques covered have to do with improving completion accuracy, and thus reliability, in particular self-consistency(@wang2022selfconsistency). However, there are a number of other techniques that can be used to improve reliability, beyond basic prompting strategies.

%%LLMs|LLM%% exhibit various problems including hallucinations(@ye2022unreliability), flawed explanations with %%CoT|CoT prompting%% methods(@ye2022unreliability), and multiple biases including majority label bias, recency bias, and common token bias(@zhao2021calibrate). Additionally, zero-shot CoT can be particularly biased when dealing with sensitive topics (@shaikh2022second).

Common solutions to some of these problems include calibrators to remove a priori biases, and verifiers to score completions, as well as promoting diversity in completions.