Recent searches

No recent searches

Search options

Only available when logged in.
mstdn.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A general-purpose Mastodon server with a 500 character limit. All languages are welcome.

Administered by:

Server stats:

11K
active users

mstdn.social: AboutStatusProfiles directoryPrivacy policyTerms of service

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.1

#MachineLearning

105 posts84 participants20 posts today
Hacker News

From Memorization to Reasoning in the Spectrum of Loss Curvature

arxiv.org/abs/2510.24256

arXiv logo
arXiv.orgFrom Memorization to Reasoning in the Spectrum of Loss CurvatureWe characterize how memorization is represented in transformer models and show that it can be disentangled in the weights of both language models (LMs) and vision transformers (ViTs) using a decomposition based on the loss landscape curvature. This insight is based on prior theoretical and empirical work showing that the curvature for memorized training points is much sharper than non memorized, meaning ordering weight components from high to low curvature can reveal a distinction without explicit labels. This motivates a weight editing procedure that suppresses far more recitation of untargeted memorized data more effectively than a recent unlearning method (BalancedSubnet), while maintaining lower perplexity. Since the basis of curvature has a natural interpretation for shared structure in model weights, we analyze the editing procedure extensively on its effect on downstream tasks in LMs, and find that fact retrieval and arithmetic are specifically and consistently negatively affected, even though open book fact retrieval and general logical reasoning is conserved. We posit these tasks rely heavily on specialized directions in weight space rather than general purpose mechanisms, regardless of whether those individual datapoints are memorized. We support this by showing a correspondence between task data's activation strength with low curvature components that we edit out, and the drop in task performance after the edit. Our work enhances the understanding of memorization in neural networks with practical applications towards removing it, and provides evidence for idiosyncratic, narrowly-used structures involved in solving tasks like math and fact retrieval.
#HackerNews#Memorization#Reasoning
TrendingLive feeds
About