Research map / Architectures and efficiency / Attention and transformers
The Impact of Positional Encoding on Length Generalization in Transformers
Amirhossein Kazemnejad, Inkit Padhi, Karthikeyan Natesan, Payel Das, Siva Reddy
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Remove explicit positional encoding (NoPE) instead of using an explicit PE | Decoder-only Transformers trained from scratch · Reasoning and mathematical downstream tasks, length extrapolation | exact-match accuracy | Improved |
| Use T5 Relative Bias instead of ALiBi, APE, or Rotary | Decoder-only Transformers · Reasoning and mathematical downstream tasks, length extrapolation | exact-match accuracy | Improved |
| Use Rotary instead of APE | Decoder-only Transformers · Downstream tasks, length extrapolation | exact-match accuracy | Improved |
| Add scratchpad | Models with the evaluated positional encodings · Addition | exact-match accuracy | Improved |
| Add scratchpad | Models with the evaluated positional encodings · Mathematical and reasoning tasks other than addition | exact-match accuracy | No clear effect |
| Use Rotary instead of NoPE or ALiBi | 1B Transformers trained on a subset of StarCoder · Language modeling at lengths beyond the training context | perplexity | Hurt |
| Use ALiBi instead of NoPE | 1B Transformers trained on a subset of StarCoder · Language modeling at lengths larger than almost twice the training context | perplexity | Improved |
Closest papers by meaning
- A Formal Framework for Understanding Length Generalization in Transformers (ICLR 2025 · similarity 0.94)
- Functional Interpolation for Relative Positions improves Long Context Transformers (ICLR 2024 · similarity 0.92)
- Exploring Transformer Extrapolation (AAAI 2024 · similarity 0.91)
- Extrapolation by Association: Length Generalization Transfer In Transformers (NeurIPS 2025 · similarity 0.91)
- The Role of Sparsity for Length Generalization in LLMs (ICML 2025 · similarity 0.90)
- What Algorithms can Transformers Learn? A Study in Length Generalization (ICLR 2024 · similarity 0.90)
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.
