Pith. sign in

The clock and the pizza: Two stories in mechanistic explanation of neural networks

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.LG 1

years

2025 1

verdicts

CONDITIONAL 1

roles

background 1

polarities

unclear 1

representative citing papers

Harmonic Loss Trains Interpretable AI Models

cs.LG · 2025-02-03 · conditional · novelty 5.0

Harmonic loss, which scores logits by inverse Euclidean distance to class prototypes with a scale-invariant normalization, yields more interpretable class centers, reduced grokking, and faster convergence across MLPs, transformers, MNIST, and a small GPT-2.

citing papers explorer

Showing 1 of 1 citing paper.

  • Harmonic Loss Trains Interpretable AI Models cs.LG · 2025-02-03 · conditional · none · ref 7

    Harmonic loss, which scores logits by inverse Euclidean distance to class prototypes with a scale-invariant normalization, yields more interpretable class centers, reduced grokking, and faster convergence across MLPs, transformers, MNIST, and a small GPT-2.