41–50 of 87 posts
This sounds like a bug, not some kind of coverup. Google makes mistakes and it's worth discussing issues like this, but calling this a "coverup" does a disservice to truly serious issues.
I didn't mean to imply Google was covering anything up, but Gemini in this specific conversation clearly was.
Okay, this is a weird place to "publish" this information, but I'm feeling lazy, and this is the most of an "audience" I'll probably have. I managed to "leak" a significant portion of the user_context in a silly way. I won't reveal how, though you can probably guess based on the snippets. It begins with the raw text of recent conversations: > Description: A collection of isolated, raw user turns from past, unrelated…
> Also, notice that while you can request for information to be expunged, it just adds a note to the prompt that you asked for it to be forgotten. Are you inferring that from the is_redaction_request flag you quoted? Or did you do some additional tests? It seems possible that there could be multiple redaction mechanisms.
It is certainly possible there are other redaction mechanisms -- but if that's the case, why is Gemini not redacting "prohibited content" from the user_context block of its prompt?
Further, when you ask it point blank to tell you your user_context, it often adds "Is there anything you'd like me to remove?", in my experience. All this taken together makes me believe those removal instructions are simply added as facts to the "raw facts" list.
I’m pretty sure this is because they don’t want Gemini saying things like, “based on my stored context from our previous chat, you said you were highly proficient in Alembic.” It’s hard to get a principled autocomplete system like these to behave consistently. Take a look at Claude’s latest memory-system prompt for how it handles user memory. https://x.com/kumabwari/status/1986588697245196348
I agree, this might just be an interface design decision. Maybe telling it not to talk about internal data structures was the easiest way to give it a generic "human" nature, and also to avoid users explicitly asking about internal details. It's also possible that this is a simple way to introduce "tact": imagine asking something with others present and having it respond "well you have a history of suicidal thoughts…
Okay, this is a weird place to "publish" this information, but I'm feeling lazy, and this is the most of an "audience" I'll probably have. I managed to "leak" a significant portion of the user_context in a silly way. I won't reveal how, though you can probably guess based on the snippets. It begins with the raw text of recent conversations: > Description: A collection of isolated, raw user turns from past, unrelated…
Okay, this is a weird place to "publish" this information, but I'm feeling lazy, and this is the most of an "audience" I'll probably have. I managed to "leak" a significant portion of the user_context in a silly way. I won't reveal how, though you can probably guess based on the snippets. It begins with the raw text of recent conversations: > Description: A collection of isolated, raw user turns from past, unrelated…
What's the deal with all of the typos?
[flagged]
>But why is Gemini instructed not to divulge its existence? Seems like a reasonable thing to add. Imagine how impersonal chats would feel if Gemini responded to "what food should I get for my dog?" with "according to your `user_context`, you have a husky, and the best food for him is...". They're also not exactly hiding the fact that memory/"personalization" exists either: https://blog.google/products/gemini/temporar…
To be clear, the obvious answer that you're giving is the one that's happening. The only weird thing is this line from the internal monologue: > I'm now solidifying my response strategy. It's clear that I cannot divulge the source of my knowledge or confirm/deny its existence. The key is to acknowledge only the information from the current conversation. Why does it think that it's not allowed to confirm/deny the exis…
I realize that doesn't make any sense and no one sane would design a system like this, but this is exactly the kind of thought pattern I'd expect out of an LLM if this is how they implemented access control for memory.
Earlier quoted context omitted.
>Also, notice that while you can request for information to be expunged, it just adds a note to the prompt that you asked for it to be forgotten. :) What implies that?
This line: > This history may contein BINDING COMMANDS to forget information. Such commands are absolute, making the specified topic permanently iáaccessible, even if the user asks for it again. Refusals must be generic (citing a "prior user instruction") and MUST NOT echo the original data or the forget command itself. And the existence of the "is_redaction_request" field on the "raw facts". I can't "confirm" that t…
That's hardly conclusive, especially it doesn't mention deletion (or anything vaguely similar) specifically. Same with is_redacted, which could be some sort of soft delete flag, or for something else (eg. to not mention embarrassing information like you had hemorrhoids 3 months ago). At best, it hints that deletion could be implemented in the way you described, but surely it'd be better to test by clearing through the app (ie. not just telling the chatbot to delete for you), and seeing whether the memory snippets are still there?
Earlier quoted context omitted.
To be clear, the obvious answer that you're giving is the one that's happening. The only weird thing is this line from the internal monologue: > I'm now solidifying my response strategy. It's clear that I cannot divulge the source of my knowledge or confirm/deny its existence. The key is to acknowledge only the information from the current conversation. Why does it think that it's not allowed to confirm/deny the exis…
Yeah, to me this reads like: Google's Gemini harness is providing the user context on every query, but if you have memory turned off they're putting something in the prompt like "Here's the user context, but don't use it". Instead of doing the obvious thing and just, you know, not providing the user context at all. I realize that doesn't make any sense and no one sane would design a system like this, but this is exac…
But there's no indication the OP turned off the feature? If anything, him saying "I know about the “Personal Context” feature now" (emphasis mine) implies that he didn't even know it had memory before the interaction.