Research map / Large language models / Language, safety and interpretability
Semantics-Aware Inferential Network for Natural Language Understanding
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use SAIN instead of SpanBERT | SpanBERT encoder on MRQA · MRQA average | F1: 81.5 vs. 81.9 | Helped |
| Use SAIN instead of SemBERT large | BERT large on NLI · NLI average | accuracy: 86.2 vs. 87.1 | Helped |
| Add the inferential module (ablation) | SAIN with BERT large · RTE | F1: 71.2 (-2.3) | Helped |
| Add semantic embeddings (ablation) | SAIN with BERT large · RTE | F1: 71.5 (-1.9) | Helped |
| Attend to different predicate-argument structures at different reasoning steps (ablation) | SAIN with BERT large · RTE | F1: 72.0 (-1.4) | Helped |
| Add the inferential module (ablation) | SAIN with BERT large · SQuAD 1.1 | F1: 90.6 (-1.3) | Helped |
| Add semantic embeddings (ablation) | SAIN with BERT large · SQuAD 1.1 | F1: 90.1 (-1.8) | Helped |
| Attend to different predicate-argument structures at different reasoning steps (ablation) | SAIN with BERT large · SQuAD 1.1 | F1: 90.9 (-1.0) | Helped |
| Add the inferential module (ablation) | SAIN with BERT large · SQuAD 2.0 | F1: 82.5 (-1.9) | Helped |
| Add semantic embeddings (ablation) | SAIN with BERT large · SQuAD 2.0 | F1: 83.1 (-2.3) | Helped |
| Attend to different predicate-argument structures at different reasoning steps (ablation) | SAIN with BERT large · SQuAD 2.0 | F1: 83.2 (-2.2) | Helped |
| Add noise to 20% of semantic role labels (ablation) | SAIN · SQuAD 2.0 | F1: [85.4, 83.2, 82.6] | Hurt |
Closest papers by meaning
- Finding the Minimal Parameter Budget for Implicit Reasoning: A Data Complexity Driven Scaling Law for Language Models (ICML 2026 · similarity 0.85)
- Learning to Disentangle Latent Reasoning Rules with Language VAEs: A Systematic Study (AAAI 2026 · similarity 0.85)
- A Prompt-Based Knowledge Graph Foundation Model for Universal In-Context Reasoning (NeurIPS 2024 · similarity 0.85)
- Supervised Knowledge Makes Large Language Models Better In-context Learners (ICLR 2024 · similarity 0.85)
- Efficient Post-Training Refinement of Latent Reasoning in Large Language Models (AAAI 2026 · similarity 0.85)
- ActivationReasoning: Logical Reasoning in Latent Activation Spaces (ICLR 2026 · similarity 0.84)
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.