Language modeling is an attempt at advancing the language intelligence of machines. These models have gained attention from researchers and companies in various domains. Hallucination in language models is a phenomenon where the output generated by models strays from factual reality or contains fabricated information. Hallucinations occlude language models from achieving their full potential in factually critical domains with higher stakes. As such, addressing hallucination is paramount, especially in domains demanding factual accuracy, such as journalism, healthcare, etc. Research has made strides in reducing hallucinations; each method suffers from its own unique imperfections. This paper aims to provide an overview of various decoding-based and related methods that aim to elucidate and mitigate the generation of hallucinatory text by language models.