cd /news/machine-learning/emergent-hierarchical-monosemantic-n… · home topics machine-learning article
[ARTICLE · art-66399] src=arxiv.org ↗ pub= topic=machine-learning verified=true sentiment=↑ positive

Emergent Hierarchical Monosemantic Neurons from the Group-Contrastive Forward-Forward Algorithm

Researchers propose Group-Contrastive Forward-Forward (GCFF), a biologically plausible learning algorithm that yields monosemantic neurons organized into hierarchies of increasing abstraction without sparsity constraints. On CLIP representations, a single trained GCFF module recovers monosemantic neurons whose abstraction increases progressively with depth, reaching environmental properties independent of an image's foreground. GCFF also achieves state-of-the-art performance among forward-forward algorithms on image classification benchmarks.

read1 min views2 publishedJul 21, 2026

arXiv:2607.16295v1 Announce Type: new Abstract: Mechanistic interpretability has made significant strides in understanding neural network representations, with sparse dictionary learning (SDL) methods, most prominently sparse autoencoders, as a central paradigm. However, recent work has reported several limitations of this paradigm: SDL objectives are non-identifiable; SDL methods rely heavily on the Linear Representation Hypothesis; and a growing body of evidence points to concepts that are encoded non-linearly and are therefore not expressible as any single direction. We hypothesise that a different route to monosemanticity is available. Biological visual systems exhibit highly selective neurons organised into hierarchies of increasing abstraction, and this organisation emerges from local, layer-wise learning rules rather than from a global error signal; we therefore ask whether a biologically plausible learning algorithm will likewise yield monosemantic neurons. To test this, we propose Group-Contrastive Forward-Forward (GCFF), a forward-forward training algorithm that combines class-specific routing with within-class contrastive objectives, reaching monosemanticity through architectural constraints rather than sparsity. Because GCFF attaches multiple non-linear layers to the representation under study, its neurons can therefore capture the non-linear concepts. On CLIP representations, a single trained GCFF module recovers monosemantic neurons whose abstraction increases progressively with depth, reaching environmental properties that hold independently of an image's foreground, without any sparsity constraint or supervision of abstraction level. We further demonstrate that GCFF can train networks from scratch, achieving state-of-the-art performance among forward-forward algorithms on various image classification benchmarks.

── more in #machine-learning 4 stories · sorted by recency
── more on @group-contrastive forward-forward 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/emergent-hierarchica…] indexed:0 read:1min 2026-07-21 ·