windowsxp811203/qwen3.8-conv1d-sigscalesync
The conv1d sigma outliers in Qwen3.8-27B are real. The quantization story attached to them is not — and the rescale variant I tested cost perplexity while fixing nothing measurable. Scope revision (2026-08-24). An earlier version of this README said the rescale was applied "exactly as published" and that the mechanism "cannot work". Both were too strong. What I tested is a 7-tensor median-normalization variant (σ > 1.5× median → α = median/σ); the intervention actually… See the full description on the dataset page: https://huggingface.co/datasets/windowsxp811203/qwen3.8-conv1d-sigscalesync.
Conversations for this repository live on Hugging Face.
CoolFace shows imported repositories read-only. Posting into someone else’s repository from here would need an authorised integration and the account holder’s consent, so the link goes to the source instead.
Open discussions on Hugging Face