undefined 3 years ago 3 comments neverrroot Reply infotainment 3 years ago As far as "leaks" go this seems a bit underwhelming. It's a general summary and not an actual prompt -- a summary that could quite easily be surmised based on the actual outputs. SwiftyBug 3 years ago I assume this is just some form of hallucination. I don't think Google would train its LLM with details of its own heuristics.
infotainment 3 years ago As far as "leaks" go this seems a bit underwhelming. It's a general summary and not an actual prompt -- a summary that could quite easily be surmised based on the actual outputs.
SwiftyBug 3 years ago I assume this is just some form of hallucination. I don't think Google would train its LLM with details of its own heuristics.
As far as "leaks" go this seems a bit underwhelming. It's a general summary and not an actual prompt -- a summary that could quite easily be surmised based on the actual outputs.
I assume this is just some form of hallucination. I don't think Google would train its LLM with details of its own heuristics.