131–140 of 246 posts
Earlier quoted context omitted.
> Three honest possibilities, and I’ll be straight with you about each This. I don't know if the "honest answer" phrasing is part of the system prompt or alignment, but when people say "honestly" all the time I start wondering how honest they're being.
At this point, I'm starting to wonder if their honesty is even load-bearing at all?
Earlier quoted context omitted.
You’re absolutely right to be hopeful. Three honest possibilities, and I’ll be straight with you about each: 1. It overthinks — Just like the previous iteration. High confidence. 2. It doesn’t overthink — Improvement from the last model for your use case. Regression for others. 3. It sometimes overthinks — Best case all around. A feature, not an impairment. One final thing worth mentioning: (I made myself irrationall…
> You’re absolutely right to be hopeful. Three honest possibilities, and I’ll be straight with you about each: > [UGC styled humorously as LLMisms] All joking aside, having interacted with Claude intensely for the last 8 months and about 30 hours/week in the last 3, I’ve started to notice how (for want of a better word) “readable” (“digestible” ? “comprehensible” ? “Predictable” is the wrong direction.) information c…
Social media killed our attention span. Now, it is being tokenized.-
> Qwen3.8-Flash-Next features a 125B-parameter main model, supplemented by an additional 51B N-gram embeddings, with 6B parameters activated per token. Didn’t see this mentioned yet. I wonder what this means for the effective size. It’s evidently ~176B paramètres, but how does that get quantized. A 4-bit quant under 100GB seems unlikely, I’m suspecting this won’t run in 128GB unified memory In principle I like the id…
The N-gram parameters can be fetched from SSD, with maybe the hottest ones staying in memory.
nVME paging the n-gram table (in BF16 for now).
Still working at it. Prefill sucks still but decode is about 12 tok/sec and the model weights fit nicely in the 128GB Spark memory in nvfp4 quant while paging the ngram stuff from disk.
(EDIT: merged to main. 80tok/sec prefill, 12 tok/sec decode, ~80GiB resident, the rest paged)
> Qwen3.8-Flash-Next features a 125B-parameter main model, supplemented by an additional 51B N-gram embeddings, with 6B parameters activated per token. Didn’t see this mentioned yet. I wonder what this means for the effective size. It’s evidently ~176B paramètres, but how does that get quantized. A 4-bit quant under 100GB seems unlikely, I’m suspecting this won’t run in 128GB unified memory In principle I like the id…
On branch of https://github.com/rdaum/eider (for DGX Spark). ~12 tok/sec decode without speculative decoding (will come later)
Still actively working on this. Prefill currently sucks. Will merge to main by end of day.
EDIT: This has now landed on main. Still haven't done MTP speculative decoding boost, but:
80tok/sec prefill, 12 tok/sec decode. ~90GiB or so resident. n-grams paged from disk.
Earlier quoted context omitted.
You’re absolutely right to be hopeful. Three honest possibilities, and I’ll be straight with you about each: 1. It overthinks — Just like the previous iteration. High confidence. 2. It doesn’t overthink — Improvement from the last model for your use case. Regression for others. 3. It sometimes overthinks — Best case all around. A feature, not an impairment. One final thing worth mentioning: (I made myself irrationall…
> Three honest possibilities, and I’ll be straight with you about each This. I don't know if the "honest answer" phrasing is part of the system prompt or alignment, but when people say "honestly" all the time I start wondering how honest they're being.
Adding to my homelab stack, hopefully it doesn't overthink like the little model. Actually, hoping it thinks a bit less. Wait actually I'm really praying it reasons a bit more directly. But wait, I'm really sure that it must be a bit better.
You might already know this, but a large part of test-time compute / 'overthinking' is just letting the model do more passes, and refine its activation residuals more. For example, even if you make thinking tokens literally just '....' (absolutely meaningless; zero information), you still see significant performance improvements: https://arxiv.org/abs/2404.15758 and https://arxiv.org/abs/2607.22925 for some starters.…
Generally speaking yes, but actually no (just randomness is suboptimal, adding steps just to add steps is suboptimal). There is a mechanism working there (in having a CoT) that is not quite clear.
The task is to optimize the efficiency of CoT. Understanding that it is not a plain "chain of thought" is the start of the problem, the solution is not there yet.
If we had the solution, there would exist no overthinking - CoT would be optimal (lean and essential plus best results).
I ran some pelicans at the four different reasoning levels (none, low, medium, xhigh - apparently high and xhigh are aliases of each other) on a DGX Spark using Unsloth's unsloth/Qwen3.8-Flash-Next-GGUF (UD-IQ1_S): https://tools.simonwillison.net/markdown-svg-renderer#url=ht... Surprised I didn't get one I liked as much as the Qwen 3.8 27B one https://simonwillison.net/2026/Aug/16/qwen-38-27b/#the-defau... , maybe be…
Adding to my homelab stack, hopefully it doesn't overthink like the little model. Actually, hoping it thinks a bit less. Wait actually I'm really praying it reasons a bit more directly. But wait, I'm really sure that it must be a bit better.
Yep.. for 'general purpose' use I found qwen3.8:27b to be disappointing due to overthinking. It's brutal especially considering how slow it is compared to MoE variants. It often overthinks to the magnitude of ~10x the tokens vs a ~4x faster gemma4:26b-a3b. As a result, qwen3.8 will churn over a prompt often for 5-10 minutes while gemma4 regularly finishes the same prompt in under 20 seconds, while giving a consistent…
Adding to my homelab stack, hopefully it doesn't overthink like the little model. Actually, hoping it thinks a bit less. Wait actually I'm really praying it reasons a bit more directly. But wait, I'm really sure that it must be a bit better.
I ran some pelicans at the four different reasoning levels (none, low, medium, xhigh - apparently high and xhigh are aliases of each other) on a DGX Spark using Unsloth's unsloth/Qwen3.8-Flash-Next-GGUF (UD-IQ1_S): https://tools.simonwillison.net/markdown-svg-renderer#url=ht... Surprised I didn't get one I liked as much as the Qwen 3.8 27B one https://simonwillison.net/2026/Aug/16/qwen-38-27b/#the-defau... , maybe be…