Paper claims LM head destroys 95-99% of training signal in backpropagation
LChoshen · x · 2026-08-17
A new paper titled "Lost in Backpropagation: The LM Head is a Gradient Bottleneck" argues that the output layer of LLMs destroys 95-99% of the training signal during backpropagation, significantly slowing down pretraining. The author discusses how a network without this bottleneck might function.
More from Research
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- GAE paper learns a geometry-native latent space for 3D-consistent world generation — yshan2u · 2026-10-16
- Agents turn conservative as context grows: edits shrink from 150 to 30 lines — dbreunig · 2026-10-16
- LMR-BENCH at EMNLP 2025 tests whether LLM agents can reproduce real LLM research — ZhiyuChen4 · 2026-10-16
- Ex-OpenAI researcher claims most neurons are legible now; critics call it hype — anshulkundaje · 2026-10-16
- Physicist pushes back on ML community's acceptance that models may be beyond understanding — tszzl · 2026-10-16