[ot][spam][log] attempt bits: talk to a language model on list

[Undescribed Horrific Abuse, One Victim & Survivor of Many]

paper information re long context:
ReRoPE [27] and LM-Infinite [13] are two other approaches that are uncompared
two compared approaches are Together.ai [31] and Code Llama [25] .

it looks to me like the together.ai approach had better performance
and smaller model size; but yarn outperformed both.