lappa
Joined 12 years ago
Karma
663
Submissions
176
Member since
Jan 25
Recent submissions
Comment · 5 months ago
ChatGPT, request minimal necessary diff to make a specific change, review, ctrl+c, ctrl+v
Comment · 8 months ago
Good post. I think you mixed torch.eye with torch.full though
Comment · 1 year ago
This isn't suggesting no one understands how these models are architected, nor is anyone saying that SDPA / matrix multiplication isn't understood by those who create these systems.
What's being said is that the result of training and the way in which information is processed in latent space is opaque.
There are strategies to dissect a models inner workings, but this is an active field of research and incomplete.
Comment · 1 year ago
I use the SingleFile extension to archive every page I visit.
It's easy to set up, but be warned, it takes up a lot of disk space.
$ du -h ~/archive/webpages
1.1T /home/andrew/archive/webpages
Comment · 1 year ago
OpenAI clearly states that they train on your data https://help.openai.com/en/articles/5722486-how-your-data-is...
Comment · 1 year ago
It's easy to argue that Llama-3.3 8B performs better than GPT-3.5. Compare their benchmarks, and try the two side-by-side.
Phi-4 is yet another step towards a small, open, GPT-4 level model. I think we're getting quite close.
Check the benchmarks comparing to GPT-4o on the first page of their technical report if you haven't already https://arxiv.org/pdf/2412.08905
Comment · 2 years ago
Great project, looking forward to seeing more as this develops.
Also FYI, your mail server seems to be down.
Comment · 2 years ago
Provided a constant temperature of 1.0, you can train the model on prompts with probablistic requests, with loss determined by KL divergence.
Expectation: 80% left, 20% right
Model sampling probability: 99% left, 1% right
>>> 0.80 * math.log(0.99 / 0.80) + 0.20 * math.log(0.01 / 0.20)
-0.42867188234223175
Model sampling probability: 90% left, 10% right
>>> 0.80 * math.log(0.9 / 0.80) + 0.20 * math.log(0.1 / 0.20)
-0.04440300758688229
Of course, if you change the temperature this will break any probablistic expectations from training in this manner.
Comment · 2 years ago
Look at it from an algorithmic perspective. In computer science many algorithms take a non-constant number of steps to execute. However, in transformers models, there are a limited number of decoder blocks, and a limited number of FFN layers in each block. This presents a theoretical upper bound on the complexity of the algorithms a decoder network can solve in a single token generation pass.
This explains why GPT4 cannot accurately perform large number multiplication and decimal exponentiation. [0]
This example can extend to general natural language generation. While some answers can be immediately retrieved or generated by a "cache" / algorithm which exists in latent space, some tokens have better quality when their latent-space algorithm is executed in multiple steps.
[0] https://www.semanticscholar.org/reader/817e52b815560f95171d8...
Comment · 2 years ago
Very interested in the expansion of RL for transformers, but I can't quite tell what this project is.
Could you please add links to the documentation to the readme where it states "It includes detailed documentation".
Also maybe DPO should use the DDPG acronym instead so your repos Deterministic Policy Optimization isn't confused for trl's Direct Preference Optimization.
Comment · 2 years ago
A few days ago I saw a post using NeuralFlow to help explain the repetition problem.
https://old.reddit.com/r/LocalLLaMA/comments/1ap8mxh/what_ca...
> I’ve done some investigation into this. In a well trained model, if you plot the intermediate output for the last token in the sequence, you see the values update gradually layer to layer. In a model that produces repeating sequences I almost always see a sudden discontinuity at some specific layer. The residual connections are basically flooding the next layer with a distribution of values outside anything else in the dataset.
> The discontinuity is pretty classic overfitting. You’ve both trained a specific token to attend primarily to itself and also incentivized that token to be sampled more often. The result is that if that token is ever included at the end of the context the model is incentivized to repeat it again.
...
> Literally just plotting the output of the layer normalized between zero and one. For one token in mistral 7B it’s a 4096 dimension tensor. Because of the residual connections if you plot that graph for every layer you get a really nice visualization.
> Edit: Here's my visualization. It’s a simple idea but I've never personally seen it done before. AFAIK this is a somewhat novel way to look at transformer layer output.
> Initial output: https://imgur.com/sMwEFEw
> Over-fit output: https://imgur.com/a0obyUj
> Second edit: Code to generate the visualization: https://github.com/valine/NeuralFlow
This is nearly identical to the overfitting example in the repo, only really representing a binary, but it's a good start. Perhaps some transformations can be applied to help further?
You can reach me at [email protected]