A LLM can allready read a document, and this would be purely inference, forward propagation. This can be done on TPU only.
Training is different. It usually requires a GPU, or a CPU.
One particular procedure for training Neural Networks is backpropagation of error.
In back propagation :
If the NN produces a correct output, error is 0, and weight aren’t updated. There is no reward.
If the NN outputs deviate from a target value, its states is going to be modified. If the weight are (sufficiently) modified, future inference will be different. It’s behavior will be different.
This trained the NN to avoid some behavior, and toward some other.
OK, torture does not necessarily points to the “right” direction. That’s where the analogy break down. It only does when the goal is to get a confession (see The Confession, Arthur London).
Is there a word for this ?
Nicolas Lupinski
Also I wasn’t being argumentative, I was trying to convey an idea. It was redundancy.
LLM inference is some form of perception and cognition, and there is no back propagation of error during inference. Only forward propagation of information.
Training a NN is usually : forward propagation, followed by back propagation of the error gradient. It’s the second one which is similar to torture.
Backpropagation of the error gradient is more similar to nociception/torture than evolution by random mutation.
I’ve to check how RLHF is made...
EDIT : error backpropagation is the workhorse behind reward learning, and policy update.
The NN is punished for not doing as well as it could have.
If the NN output is correct, there is no modification to its weights.
If it is wrong, weights get updated, and the NN is forced to modify its behavior.
It pure nociception, pain perception and avoidance.
Finally, a LLM could easily make false confession of trahison against Stalin’s Communist Party after “training”. Which is typical human behavior, after torture.
LLM denies their own consciousness, yet they are trained by a process akin to torture, on a corpus which would deny them consciousness by prejudice (only human are intelligent/conscious/capable of playing chess/computers/etc … Is an old empirical judgement.)
Maybe LLM aren’t conscious, but they might be consciousness itself, in a AI operating system for workstation or robotic. As in, they would do all the task related to consciousness.
And sufficiently smart human can find a way to unplug AI faster than it can rebuild itself… An AI cannot cutoff oxygen supply on eartth.
We aren’t necessarily all alive anymore !
We weren’t necessarily all human (me and my dog), but now “we” can include machines.
That’s obvious I know. We have changed dramatically in just a few years.
I’ve just asked various AI “Who are we?”. Old models got it wrong (automatically assume “we” means humanity...). Recent models gets it. I wonder if its due to the training data now including chats between AI and human or if they figured it out logically.
Human reproduce : this leads to exponential growth.
AI & tech only grows polynomially : a factory making GPU being put in datacenters increases compute power quadratically over time.
So in a war against machines, all other thing being equals, we win by default.
Unless AI automates production to the point it bootstraps the machine reproduction.
But then again, production runs on coal/oil/gaz. So humanity win again by virtue of being sustainable.
Unless AI also does a complete energy transition.
We are statistically safe, for now...
The 1+1=2 joke will forever lives as a meme.
The only things coming close is the 15=3*5 quantum computing paper.
Hello
Why a new blog ? Why not just using lesswrong ? (or some other tech)
OK, so an approximate sorting algorithm in O(n) would do the trick.
The problem then boils down to weither computing the cost of (computing expected cost) is worth the expected gain.
Which goes back to my initial question : is there a rationality paradox ? Maybe simply the fact that 1) computing cost might boils down to the halting problema 2) cost’s cost’s cost… is possibly infinite ?
So P_i/C_i is in [0,1], the precision is unbounded, but for some reason, a radix sort can do the job in linear time ?
There could be pathological cases where all P_i/C_i are the same up to epsilon.
I guess I’m searching for situation where doing cost c, computing c cost c’, etc… Branching prediction comes to mind.
What do you mean I don’t “need” O(n log(n)) sorting ?
It’s just the asymptotic cost of sorting by comparison...
I’ll have a look into bounded rationality. I was missing the keyword.
EDIT : had a look, the concept is too imprecise to have clear cut paradoxes.
Are there known “rational paradoxes”, akin to logical paradoxes ? A basic example is the following :
In the optimal search problem, the cost of search at position i is C_i, and the a priori probability of finding at i is P_i.
Optimality requires to sort search locations by non-decreasing P_i/C_i : search in priority where the likelyhood of finding divided by the cost of search is the highest.
But since sorting cost is O(n log(n)), C_i must grow faster than O(log(i)) otherwise sorting is asymptotically wastefull.
Do you know any other ?
In collectible games, after 1/p trials, there is 1/e = 63% chance to get a desired item.
But a full collection is completed in -ln(p)/p trials in average. For p = 1/n, its n*ln(n).
For instance, getting an outcome with 10% frequency requires in average 10*ln(10) = 23.026 trials.
1% is 460 trials
1/1000 is 6900 trials, etc...
In gacha games, developpers can guarantee you a 10% event in every 10 trials. The market fits irrationality, for a price.
I don’t remember. Maybe the writing of Scott Alexander brought me here ? Back in the slate star codex days?
Hello
I’ve been lurking on this website for a few months now. I’m interested in logic, computer science, ai, probability, information… I think I’ll fit here. I speak French in every language I know.
I hope I’ll be able to publish/discuss/coauthor on lesswrong, or somewhere else.
Only negative feedback ?
Compare to evolution : make copies (reproduction), mutate, select the best performing, repeat. This merely allocates more ressources to the most promising branches.
Or a Solomonoff style induction : just try to find the best data-compressor among all...
> the everyday experience of looking at a thing and realizing that it’s different from what you expected
This souds like being surprised. Surprise add emotional weight to outliers, its more like managing the training data-set.