The job of the future.
700 karma · joined October 16, 2023
The performance changes are so big with the right harness that is makes sense to engineer the harness and fine-tune the model to one another from the start.
• Terminal Bench: 56.9 → 82.7 (+25.8)
• Toolathlon: 51.8 → 70.3 (+18.5)
Compared to GPT-5.6 Terra:
• Terminal Bench: Flash 82.7 vs Terra 78.4
• Toolathlon: Flash 70.3 vs Terra 53.1
• DeepSWE: Flash 54.4 vs Terra 69.6
• Agents' Last Exam: Flash 25.2 vs Terra 50.4
Trading blows with Terra, which is pretty interesting. No clear winner on these benchmarks, and wildy differeing scores. Very interesting!
Yeah, the encoder and decoder stuff is explicit, but the internal structure in generated during training. I don't think the big labs were doing this back when I did the research; no one was back in '24.
I just didn't get round to publishing for years, because I have a day job.
By the way, it still works! I tested it earlier this year on Qen3.6 and you still see improvements, so either a) no one actually paid attention, or b) it has more room to scale.
curl -X POST https://quantumlever.stream/api/magic-8-ball \ -H "Content-Type: application/json" \ -d '{"question":"Will my sampler touch the multiverse?"}'
Author here.
I drafted it before I left for holiday, at it's not ready to publish.
It wasn't supposed to be officially posted yet, but I ran out of time before my flight.
My apologies!
I have pushed basic code to GitHub (https://github.com/dnhkng/RYS)
Some interesting areas to explore might be a combination of deleting some layers and duplicating others. i.e. reduce VRAM by dropping some layer (this works, well documented), and recovering performance by duplicating others (saves VRAM). I am not pursuing this, but it seems interesting!
The probes I used seem to help identify good configurations, but are quite noisey. A small probe set was initially used to make the scan tractable, and then the higher ranked models were retested on a set ~10x larger.
It turns out this does not help (somewhat surprisingly).
I am working with TurboDerp to integrate this into the Exllama v3 format.
I extended this to a 2×2 design (two languages × two content types) and the result is even starker: by layer 10, cross-language same-content pairs are more similar than same-language different-content pairs. The model cares about what you're saying, not what language you're saying it in.
This is also what makes layer duplication work — those mid-stack layers operate in a space where input and output distributions match, so you can loop through them without breaking anything. The encoding and decoding boundaries are where the blue walls show up in the heatmaps.
I think this says something interesting about how transformers organise computation internally. The mid-stack reasoning circuits are coherent enough that you can loop through them twice without distribution mismatch. The encoding/decoding boundaries are not.
I have Deepseek etc, but inferencing on DDR5 would take about 2-3 weeks for a simple scan. I think this works best with dense models, but it also seems ok with MoE.
@everyone: Can someone hook me up with Nvidia sponsorship?
I do wish one of the big labs would sponsor with a rack of HGX Rubin NVL8's. I have lots of ideas to test, and I have probably hit the spending limit with the boss on hardware (she hasn't seen the new power bill yet...)
But that points again to the main idea: The model has learnt to transform Base64 into a form it can already use in the 'regular' thinking structures.
The alternative is that there is an entire parallel structure just for Base64, which based on my 'chats' with LLMs in that format seems implausible; it acts like the regular model.
If there is a 'translation' organ in the model, why not a math or emotion processing organs? Thats what I set out to find, and are illustrated in the heatmaps.
Also, any writing tips from the Master blogger himself? Huge fan (squeal!)
On the other papers, models like SOLAR or training a model that uses a single layers are probably going to hit a wall, based on the heatmaps I found. The transformer stack starts with randomised weights, (analogous to undifferentiated stem cells), and it seems they later form 'organs' during the trillions of pre-training tokens they undergo. My hypothesis is that you probably only want one copy of the 'token-to-thought', and 'thought-to-token' organs. It seems that you can make one layer do all three things (transforms in and out, and do the 'thinking'), but I think specialisation will always win.
On your questions: 1) A few other papers have been mentioned in the thread, like Solar10.7B. They duplicated the whole transformer stack, and it kinda helped. But as I found experimentally, that probably not a great idea. You are duplicating 'organs' (i.e. input processing stuff), that should only have one copy. Also, that paper didn't see immediate improvements; they had to do continued pre-training to see benefits. At that point, I'm guessing the big labs stopped bothering. Limited by hardware, I had to find unusual angles to approach this topic.
2) Nah, no more wetware for me. I did a half decade of research at a big neurobiology institute, and while it was very enjoyable, I can truly say that grant writing and paper review are 'not my thing'. This reason this info was delayed so long is that I wanted a paper in the AI field to go along with my papers in other fields. But as a Hobbyist with no official affiliation, and the attention span of a gnat, I gave up and started a blog instead. Maybe someone will cite it?
The code in the blog helps derive useful metrics from partial answers.
You can enter the setting, and apply new re-layering architectures. Its very weird chatting with these brain-damaged models.
I tried that pretty early on, the its basically never good. Its described in the the section: https://dnhkng.github.io/posts/rys/#the-beginning-of-llm-neu...