It’s amazing at debugging too. I had it running in Powershell controlling an lldb session in MSYS2. The way it can read addresses and so root cause analysis is amazing! It takes a lot of mind power to do those things.
I like to watch it work though because it honestly teaches me some tricks.
This type of project looks extremely useful. There was a lot of buzz around Jev, but having models that run locally and can be fine-tuned is extremely helpful.
Many people knew this announcement was coming. The betting markets suggested a very high likelihood of Opus dropping today. I was anticipating this quite a bit!
I just checked that model (ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S) and it does much much worse on the "Please recite Jabberwocky" test than the original bf16 does.
The bf16 only misses "snicker-snack" and this quantization becomes confused after the first stanza.
I think it should be expected that a smaller model is worse at reciting memorized data than a big one. I also don't think it's a good use case of small local models. Can it find and recite Jabberwocky if given access to a web search tool?
It depends on what you mean by lossless. If both models can perform the same tasks it can be considered lossless for those tasks. That task might be more related to language understanding rather than memorizing, they have several benchmarks in the article.
I'm going to stop here and be direct: I'm having trouble recalling the exact text, and every attempt above is me guessing. Rather than present a mangled version as the real poem, I'd recommend you look it up — it's very short and in the public domain, so any text of Through the Looking-Glass will have it verbatim. If you'd like, I can help with the moral of the poem ("'twas the blessing of the Bird..."), the famous Humpty Dumpty word interpretations ("slithy" = lithe + sinister, "mimsy" = miserable + mys... etc.), or Carroll's original annotations for the coined words — that part I can do reliably.
The benchmarks they chose are rather cherry picked to not include long context or difficult ones that involve long horizon work or many agent turns, as I suspect this is where the model shows more differences compared to the full fat one
Yep more hops from the lower Q is likely going to skew the vectors further over time.
I wonder if there's a way to mitigate this by running it through an original Q8 draft model, attuned somehow for the PTQ1 quant, but giving it a higher threshold for the acceptance linear with the context length itself?
The longer the context, the higher the multiplier on the threshold, and more likely the draft result is used. Not ideal but it may extend the usable max context.
This model might, even without this, be amazing for short lived agents that work via generations / have changing tasks.
Came to ask the same. From my really rough understanding, it seems like Unsloth's method allows a slightly higher precision at a higher file size, while PrismML's uses a different approach to achieve a smaller size (and presumably less precision).
reply