AI Paper Atlas

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

Measuring Systematic Generalization in Neural Proof Generation with Transformers

Nicolas Gontier, Koustuv Sinha, Siva Reddy, Chris Pal

NeurIPS 2020 · Large language models › Language, safety and interpretability · approach: Faithfulness of LLM explanations · 38 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (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
train on backward-chaining (reversed) proofs instead of forward-chaining proofs (ablation)2.5M Transformer decoder · CLUTRR facts and amt storiesanswer accuracyImproved
use long exhaustive proofs instead of short proofs (ablation)2.5M Transformer decoder · CLUTRRanswer accuracyImproved
generate a proof before the answer vs no-proof (answer directly) (ablation)2.5M Transformer decoder · CLUTRR facts storiesanswer accuracyHurt
forward-chaining vs backward-chaining proofs (ablation)2.5M Transformer decoder · CLUTRRproof validityImproved
provide ground-truth proof in the input (ablation)Transformer decoder · CLUTRR inductive and interpolation levelsanswer accuracyImproved
extrapolate to unseen harder levels (7-10) even when proof is givenTransformer decoder · CLUTRR extrapolationanswer accuracyHurt
train on noisy amt-template stories instead of facts-templateTransformer decoder · CLUTRRanswer accuracy and proof validityHurt
scale model from 2.5M to 145M parameters (ablation)Transformer decoder · CLUTRRanswer accuracyNo clear effect

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