atlas

Research map / Large language models / Language, safety and interpretability

Sub-Task Decomposition Enables Learning in Sequence to Sequence Tasks

ICLR 2023 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 3 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add intermediate sub-task labels to the input and train on the sub-task outputs as well as the final answerSequence-to-sequence learning of polynomial-time tasks · Tasks in Pzero-one lossHelped
Add sub-task decomposition supervisionBERT-base Transformer · Bit-subset parity, 32 bitsvalidation accuracyHelped
Train without intermediate sub-task supervisionBERT-base Transformer · Bit-subset parity, 32 bitsvalidation accuracyHurt
Remove intermediate sub-task supervisionPolynomial-time learnable task families under the stated hardness assumptions · Binary classification taskszero-one lossHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy