Locate-then-edit succeeds at the same early-to-mid MLP locations in masked diffusion models as in autoregressive models, but requires optimization over intermediate partial-mask states to handle multi-token targets.
In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 16801–16819, Miami, Florida, USA
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 3roles
background 1polarities
background 1representative citing papers
CLaRE quantifies representational entanglement in LLMs using single-layer forward activations to predict editing ripple effects, reporting 62.2% higher Spearman correlation than baselines while using 2.74x less time and 2.85x less GPU memory.
A controlled benchmark on 2040 problems reveals poor generalization and high interference in model editing for API updates in code LLMs, with many successes being workarounds rather than true migrations.
citing papers explorer
-
Knowledge Editing in Masked Diffusion Language Models
Locate-then-edit succeeds at the same early-to-mid MLP locations in masked diffusion models as in autoregressive models, but requires optimization over intermediate partial-mask states to handle multi-token targets.
-
CLaRE-ty Amid Chaos: Quantifying Representational Entanglement to Predict Ripple Effects in LLM Editing
CLaRE quantifies representational entanglement in LLMs using single-layer forward activations to predict editing ripple effects, reporting 62.2% higher Spearman correlation than baselines while using 2.74x less time and 2.85x less GPU memory.
-
Understanding Robustness of Model Editing in Code LLMs
A controlled benchmark on 2040 problems reveals poor generalization and high interference in model editing for API updates in code LLMs, with many successes being workarounds rather than true migrations.