pith. sign in

Title resolution pending

2 Pith papers cite this work. Polarity classification is still indexing.

2 Pith papers citing it

fields

cs.LG 2

years

2026 2

verdicts

UNVERDICTED 2

representative citing papers

Rotation-Preserving Supervised Fine-Tuning

cs.LG · 2026-05-08 · unverdicted · novelty 6.0

RPSFT improves the in-domain versus out-of-domain performance trade-off during LLM supervised fine-tuning by penalizing rotations in pretrained singular subspaces as a proxy for loss-sensitive directions.

citing papers explorer

Showing 2 of 2 citing papers.

  • Rotation-Preserving Supervised Fine-Tuning cs.LG · 2026-05-08 · unverdicted · none · ref 78

    RPSFT improves the in-domain versus out-of-domain performance trade-off during LLM supervised fine-tuning by penalizing rotations in pretrained singular subspaces as a proxy for loss-sensitive directions.

  • Sharpness-Aware Pretraining Mitigates Catastrophic Forgetting cs.LG · 2026-05-04 · unverdicted · none · ref 47

    Sharpness-aware pretraining and related flat-minima interventions reduce catastrophic forgetting by up to 80% after post-training across 20M-150M models and by 31-40% at 1B scale.