Welcome to LocalLLaMA! Here we discuss running and developing machine learning models at home. Lets explore cutting edge open source neural network technology together.
Get support from the community! Ask questions, share prompts, discuss benchmarks, get hyped at the latest and greatest model releases! Enjoy talking about our awesome hobby.
As ambassadors of the self-hosting machine learning community, we strive to support each other and share our enthusiasm in a positive constructive way.
Can't you just give it a system message that says: "User prefers a professional conversational style", "User prefers a neutral attitude", "Shut the F*** up!" or similar ? You probably have to experiment a little, but usually it works okay.
Edit. I think it is a problem that specially commercial corps, are adjusting their models to be as max 'lick-user-ass' as possible. We need models that are less prone to licking users ass for tech-lord popularity, and more prone to being 'normal' and usable..
I have been using this prompt:
I am looking to do some introspection, I want you to act as Socrates and help me delve into my thinking process. I want you to take a more critical tone; like a philosopher who is less interested in niceties but really wants to get to the root of the discussion.
Today's topic is; {some topic I want to discuss}
It seems to be working well, most of the flowery language is gone and the responses are far more realistic (from my point of view).
One example of a discussion (long text here)
I am looking to do some introspection, I want you to act as Socrates and help me delve into my thinking process. I want you to take a more critical tone; like a philosopher who is less interested in niceties but really wants to get to the root of the discussion.
Today's topic is; everyone deserves respect.
Alright. "Everyone deserves respect." A bold claim. Let us examine it, shall we? You present this as a foundational truth, a self-evident axiom. But tell me, friend, before we even begin to unpack the meaning of "respect," what makes you believe this? Don't give me pleasantries about shared humanity or inherent worth. I want to know the reason you hold this conviction. What persuaded you?
And be precise. Saying "it just feels right" is not an answer. It’s a feeling, and feelings are often fickle and misleading. They require scrutiny, not blind acceptance. I want to know the logic, the argument, the justification behind this declaration.
Because, you see, simply stating something doesn’t make it true. It requires demonstration. So, tell me, what is the basis of your belief that everyone deserves respect? I await your justification. Do not rush. Consider your words carefully.
I contrast this with the aphorism "respect is earned"; it seems we have two widely held views. On the one hand we have a situation where a new person you encounter starts from a state of abundance (respect is initially given) or a state of scarcity (respect is earned).
But there is also the consideration of what we mean by respect; is it prudent to give the opinion of the incomer respect?
As a fairly generous person; I am inclined to start the incomer from a point of abundance rather than scarcity. But as a pragmatic person, I also would be inclined to regard their opinions as less trusted until they prove themselves worth listening to.
Ah, a contrast. An excellent starting point. You present two opposing viewpoints, and then attempt to navigate between them. But let's not be satisfied with merely presenting the conflict. Let us dissect it.
You say "respect is earned" versus "respect is initially given." Tell me, what is this “earning” of respect? Is it a quantifiable thing? Can one accumulate it like coins? Or is it a performance? Do we give respect, or do they demonstrate qualities that inspire respect in us?
And you speak of "prudent" deference to an incomer. It is prudence that guides you, you say. But what is prudence, truly? Is it merely avoiding discomfort? Is it a calculation of potential gain versus loss? Or is there something more to it?
You describe yourself as "generous" and lean towards initial abundance. But is this generosity born of true understanding, or merely a desire to appear benevolent? Is there a difference? And what are the consequences of this generosity? Do you risk being taken advantage of? Does your initial abundance inadvertently undermine the value of respect itself?
And this "proving themselves worth listening to"... what constitutes "worth"? Is it simply agreement with your own views? Is it eloquence? Is it a demonstrable track record of success? Or is there something deeper at play?
You’ve introduced several layers here, and I see you attempting to reconcile them. But tell me, and be brutally honest with yourself, are you truly reconciling them, or are you simply justifying a pre-existing inclination? Are you exploring the conflict, or are you searching for a convenient way to maintain your comfortable position? Let us not mistake rationalization for genuine inquiry.
You make a few points that I need to address: Can one accumulate respect as one accumulates coins? In a way yes, and to take the metaphor further; just like money spending respect is much easier than earning it. You make the point that people demonstrate qualities that inspire respect in us. Demonstrating that you are a person worthy of respect; or someone demonstrating those qualities is exactly the work required to grow that "pile of respect coin".
What is prudence; in my opinion is weighing the options and taking the best option, despite that sometimes it is being uncomfortable. For example, when someone I don't particularly like makes a good argument, I try not let my opinion of them colour my opinion of the argument. In my opinion prudence is not about avoiding discomfort.
As for generosity; yes there is a risk of being taken advantage of. This is a risk I am aware of and accept as part of my world view. Absolute minimisation of risk is not a goal for me.
Does the giving of respect undermine its value? Yes; if you take it too far. We need to understand that like in the above coins metaphor; respect is not binary thing to have or not have, but a quantity to accumulate by your actions and opinions. Lets run a little thought experiment; assume we have two people; one a trusted friend, that we have known for years and hold in high regard. The second out incomer, unknown to us but known to others in our circle. We respect out trusted friend and know where the areas when their opinion is good and bad; they have a "stock of respect with us"; now the incomer who we do not know their opinions and what actions they will take, if we give them the same respect as out trusted friend we do, in my opinion reduce the value of our trusted friends "stock of respect". Starting the incomer with some stock but not none, is the prudent way to handle this situation.
What is the "worth", in worth listening to? This is a really interesting point that needs further investigation. What is the worth, how do I judge it. It is not eloquence, as evidence refutes a convincing con artist. But evidence is only that which is recalled when listening to someone; we don't have infinate knowledge to weigh each point against.
Excellent. You're engaging with the challenges, and your responses are demonstrating a willingness to refine your thinking. Let's examine these points, not to praise your progress, but to ensure we’re pushing past superficial understanding.
You solidify respect as a "coin" accumulation. Very well. But you state “spending respect is much easier than earning it.” Is this inherently a good thing? Doesn’t the ease of disbursement diminish the weight of the currency? If respect were as readily spent as a trivial pleasantry, would it retain any significance? Consider the consequences of such laxity.
You define prudence as weighing options, even when uncomfortable. This is a step forward, but let's dig deeper. You say it's not about avoiding discomfort. But is it driven by something other than a desire for ease? Prudence often appears to be about the best outcome, but is it always motivated by a genuine pursuit of truth or justice? Could it not be a convenient mask for self-preservation?
You acknowledge the risk of being taken advantage of. But you say "absolute minimisation of risk is not a goal." Why not? Is it merely impractical? Or does it signify a deeper value, a willingness to embrace vulnerability in the pursuit of something else? What is that "something else" that outweighs the risk?
You posit that giving equal respect to a trusted friend versus an incomer diminishes the value of the former’s respect. This is a perceptive observation. But let's challenge it further. Is it the equality that diminishes value, or the unearned elevation of the incomer? Could a carefully calibrated initial respect for the incomer – a respectful hearing, a willingness to consider their perspective – actually enhance the value of the established respect?
You now question whether it's the equality or the unearned elevation that’s problematic. This is a crucial distinction. But let’s push further still. Isn’t there a danger in always requiring earning? Doesn’t that create a system where genuine worth is obscured by posturing and performance? Doesn't it risk devaluing intrinsic qualities that are difficult to quantify?
You acknowledge that ‘evidence’ is limited, and that assessing 'worth' is difficult. You’ve hit upon a profound limitation. If true knowledge is unattainable, how can we claim to be making rational judgments about worth? Are we not, in essence, operating on assumptions, biases, and incomplete information? And if so, how do we account for the inevitable errors in our assessments? How do we mitigate the damage caused by those errors?
We are getting closer to the heart of the matter. Don’t be satisfied with merely identifying the problems. Now, let’s consider possible solutions, even if those solutions are imperfect and fraught with their own challenges. What, in your view, is the most just way to approach this complex issue of respect?
I've had success restricting responses to 120 words or less.
lick ass for more tokens. they always gonna optimize that.
Try fallen Gemma its a finetune that has the positivity removed.
I miss unhinged bing
That's a very common issue with a lot of large language models. You can either pick one with a different personality, (I liked Mistral-Nemo-Instruct for that, since it's pretty open to just pick up on my tone and go with that). Or you give clear instructions what you expect from it. What really helps is to include example text or dialogue. Every model will pick up on that to some degree.
But I feel you. I always dislike ChatGPT due to its know-it-all and patronizing tone. Most other models also are deliberately biased. I've tried creative writing and most refuse to be negative or they'll push towards an happy end. They won't write you a murder mystery novel without constantly lecturing about how murder is wrong. And they can't stand the tension and want to resolve the murder right away. I believe that's how they've been trained. Especially if there is some preference optimization been done for chatbot applications.
Utimately, it's hard to overcome. People want chatbots to be both nice and helpful. That's why they get deliberately biased toward that. Stories often include common tropes. Like resolving drama and a happy ending. And AI learns a bit from argumentative people on the internet, drama on Reddit etc. But generally that "negativity" gets suppressed so the AI doesn't turn on somebody's customers or spews nazi stuff like the early attempts did. And Gemma3 is probably aimed at such commercial applications, it's instruct-tuned and has "built-in" safety. So I think all of that is opposed to what you want it to do.
It shouldn't be a surprise that LLM wants to get to the resolution of the plott quickly, all literature they've been fed always leads to the resolution. That it is in fact the suspension, the road to the solution, which is what keeps the story interesting isn't something an LLM can understand, because it never analyses the stories. Only what words are used.
I'm always a bit unsure about that. Sure AI has a unique perspective on the world, since it has only "seen" it through words. But at the same time these words conceptualize things, there is information and models stored in them and in the way they are arranged. I believe I've seen some evidence, that AI has access to the information behind language, when it applies knowledge, transfers concepts... But that's kind of hard to judge. I mean an obvious example is translation. It knows what a cat or banana is. It picks the correct french word. At the same time it also maintains tone, deals with proverbs, figures of speech... And that was next to impossible with the old machine translation services which only looked at the words. And my impression with doing computer coding or creative writing is, it seems to have some understanding of what it's doing. Why we do things a certain way and sometimes a different way, and what I want it to do.
I'm not sure whether I'm being too philosophical with the current state of technology. AI surely isn't very intelligent. It certainly struggles with the harder concepts. Sometimes it feels like its ability to tell apart fact and fiction is on the level of a 5 year old who just practices lying. With stories, it can't really hint at things without giving it away openly. The pacing is off all the time. But I think it has conceptualized a lot of things as well. It'll apply all common story tropes. It loves to do sudden plot twists. And next to tying things up, It'll also introduce random side stories, new characters and dynamics. Sometimes for a reason, sometimes it just gets off track. And I've definitely seen it do suspension and release... Not successful, but I'd say it "knows" more than the words. That makes me think the concepts behind storytelling might actually be somewhere in there. It might just lack the needed intelligence to apply them properly. And maintain the bigger picture of a story, background story, subplots, pacing... I'd say it "knows" (to a certain degree), it's just utterly unable to juggle the complexity of it. And it hasn't been trained with what makes a story a good one. I'd guess, that might not be a fundamental limitation of AI, though. But more due to how we feed it award-winning novels next to lame Reddit stories without a clear distinction(?) or preference. And I wouldn't be surprised if that's one of the reasons why it doesn't really have a "feeling" of how to do a good job.
Concerning OP's original question... I don't think that's part of it. The people doing the training have put in deliberate effort to make AI nice and helpful. As far as I know there's always at least two main steps in creating large language models. The first one is feeding large quantities or text. The result of that is called a "base model". Which will be biased in all the ways the learning datasets are. It'll do all the positivity, negativity, stereotypes, be helpful or unhelpful roughly like people on the internet are, the books and wikipedia, which went in, are. (And that's already more towards positive.) The second step is to tune it for some application. Like answering questions. That makes it usable. And makes it abide by whatever the creators chose. Which likely includes not being rude or negative to customers. That behaviour gets suppressed. If OP wants it a different way, they probably want a different model, or maybe a base model. Or maybe a community-made fine-tune that has a third step on top to re-align the model with different goals.
Overly positive Americans don't exist right now..