๐ฎ
๐ฎ
The Ethereal
Which Decisions Low-Bit Quantization Breaks, and How to Predict Them
August 06, 2026 ยท Grace Period ยท ๐ EMNLP 2026
Authors
Zekun Wu, Swati Dhiman, Adriano Koshiyama
arXiv ID
2608.06564
Category
cs.LG: Machine Learning
Cross-listed
cs.CL
Citations
0
Venue
EMNLP 2026
Abstract
Quantization is how large language models are actually deployed, and below four bits it hurts. What nobody can say is which decisions change at a given bit-width -- which matters most where a model acts rather than answers, since a tool call it declines to make is a failure no score reports. A compressed agent stops calling its tools, then loses half its safety refusals, while benchmark scores barely move. Prior work assumes the added noise has a roughly fixed size, which would make confident decisions safe. We measure the decision instead: the margin between the option a model picks and its best alternative, before and after quantization, across 16 models, three methods, and 8 down to 2 bits. Kinds of decision do not break together -- at 3 bits the decision to call a tool collapses toward inaction while the choice of which tool is untouched -- and the damage is proportional rather than fixed, the margin multiplied by a factor that collapses with bit-width (median 0.86 at 4 bits, 0.33 at 3, 0.00 at 2). Fitted against additive competitors, including one whose noise grows with the margin, no account with an additive mean wins a damaged tool or safety cell; that is the best description among those stated, not a proof of generative form. Given a condition's own constants the relation predicts flip rates on its held-out decisions to a median of 1.8 points, with calibrated per-decision probabilities (calibration error 0.004 over 131,758 predictions), and no flip was used in the fit. Borrowed constants are wrong by 18-33 points at 3 bits, so a small paired margin set measured per model is the instrument, not a way to skip measuring. It is anchored to behaviour where used: at 4 bits the most likely token over the vocabulary is one of the two options in 85% of tool items, and the 2-bit floor is where the instrument stops measuring. Nothing repairs the damage more cheaply than one more bit.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
๐ Similar Papers
In the same crypt โ Machine Learning
๐ฎ
๐ฎ
The Ethereal
Continuous control with deep reinforcement learning
๐
๐
Old Age
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
๐
๐
Old Age
Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
๐
๐
Old Age
SGDR: Stochastic Gradient Descent with Warm Restarts
๐ฎ
๐ฎ
The Ethereal