Is there a way to get the low level request id and actual llm response?

[richtong]

Trying to debug sambanova which returns a 400 error. I can't figure out where a log of what's exactly send an received is so I can help them debug it.

Same problem with a.ai glm 4.6. Have a good contact there and they need a request id for queries that are wrong so they can trace.

Is there a log like this somewhere deep in too code?