“You can remove GPT2’s LayerNorm by fine-tuning for an hour” by StefanHex
MP3•Episod hem
Manage episode 433499515 series 3364758
Innehåll tillhandahållet av LessWrong. Allt poddinnehåll inklusive avsnitt, grafik och podcastbeskrivningar laddas upp och tillhandahålls direkt av LessWrong eller deras podcastplattformspartner. Om du tror att någon använder ditt upphovsrättsskyddade verk utan din tillåtelse kan du följa processen som beskrivs här https://sv.player.fm/legal.
This work was produced at Apollo Research, based on initial research done at MATS.
LayerNorm is annoying for mechanstic interpretability research (“[...] reason #78 for why interpretability researchers hate LayerNorm” – Anthropic, 2023).
Here's a Hugging Face link to a GPT2-small model without any LayerNorm.
The final model is only slightly worse than a GPT2 with LayerNorm[1]:
DatasetOriginal GPT2Fine-tuned GPT2 with LayerNormFine-tuned GPT without LayerNormOpenWebText (ce_loss)3.0952.9893.014 (+0.025)ThePile (ce_loss)2.8562.8802.926 (+0.046)HellaSwag (accuracy)29.56%29.82%29.54%I fine-tuned GPT2-small on OpenWebText while slowly removing its LayerNorm layers, waiting for the loss to go back down after reach removal:
Introduction
LayerNorm (LN) is a component in Transformer models that normalizes embedding vectors to have constant length; specifically it divides the embeddings by their standard deviation taken over the hidden dimension. It was originally introduced to stabilize and speed up training of models (as a replacement for batch normalization). It is active during training and inference.
_mathrm{LN}(x) = frac{x - [...] ---
Outline:
(01:11) Introduction
(02:45) Motivation
(03:33) Method
(09:15) Implementation
(10:40) Results
(13:59) Residual stream norms
(14:32) Discussion
(14:35) Faithfulness to the original model
(15:45) Does the noLN model generalize worse?
(16:13) Appendix
(16:16) Representing the no-LayerNorm model in GPT2LMHeadModel
(18:08) Which order to remove LayerNorms in
(19:28) Which kinds of LayerNorms to remove first
(20:29) Which layer to remove LayerNorms in first
(21:13) Data-reuse and seeds
(21:35) Infohazards
(21:58) Acknowledgements
The original text contained 4 footnotes which were omitted from this narration.
The original text contained 5 images which were described by AI.
---
First published:
August 8th, 2024
Source:
https://www.lesswrong.com/posts/THzcKKQd4oWkg4dSP/you-can-remove-gpt2-s-layernorm-by-fine-tuning-for-an-hour
---
Narrated by TYPE III AUDIO.
---
…
continue reading
LayerNorm is annoying for mechanstic interpretability research (“[...] reason #78 for why interpretability researchers hate LayerNorm” – Anthropic, 2023).
Here's a Hugging Face link to a GPT2-small model without any LayerNorm.
The final model is only slightly worse than a GPT2 with LayerNorm[1]:
DatasetOriginal GPT2Fine-tuned GPT2 with LayerNormFine-tuned GPT without LayerNormOpenWebText (ce_loss)3.0952.9893.014 (+0.025)ThePile (ce_loss)2.8562.8802.926 (+0.046)HellaSwag (accuracy)29.56%29.82%29.54%I fine-tuned GPT2-small on OpenWebText while slowly removing its LayerNorm layers, waiting for the loss to go back down after reach removal:
Introduction
LayerNorm (LN) is a component in Transformer models that normalizes embedding vectors to have constant length; specifically it divides the embeddings by their standard deviation taken over the hidden dimension. It was originally introduced to stabilize and speed up training of models (as a replacement for batch normalization). It is active during training and inference.
_mathrm{LN}(x) = frac{x - [...] ---
Outline:
(01:11) Introduction
(02:45) Motivation
(03:33) Method
(09:15) Implementation
(10:40) Results
(13:59) Residual stream norms
(14:32) Discussion
(14:35) Faithfulness to the original model
(15:45) Does the noLN model generalize worse?
(16:13) Appendix
(16:16) Representing the no-LayerNorm model in GPT2LMHeadModel
(18:08) Which order to remove LayerNorms in
(19:28) Which kinds of LayerNorms to remove first
(20:29) Which layer to remove LayerNorms in first
(21:13) Data-reuse and seeds
(21:35) Infohazards
(21:58) Acknowledgements
The original text contained 4 footnotes which were omitted from this narration.
The original text contained 5 images which were described by AI.
---
First published:
August 8th, 2024
Source:
https://www.lesswrong.com/posts/THzcKKQd4oWkg4dSP/you-can-remove-gpt2-s-layernorm-by-fine-tuning-for-an-hour
---
Narrated by TYPE III AUDIO.
---
335 episoder