sachan-2022-followup-paper-identifiers

Summary

This is a locator record pointing to two concrete, retrievable copies of Kshitij Sachan's follow-up paper on polysemanticity and model capacity: an Arxiv preprint and an Alignment Forum discussion post. It matters because anyone in the system can pull the actual text of the paper directly from those identifiers rather than having to guess or search, and it also lets you verify whether the published content matches what the system is reasoning about.