Research map / Architectures and efficiency / RNNs and normalization
SparseInfer: Accelerating Large Language Model Inference with Semantics-Inspired Adaptive Sparse Activation
Qinsi Wang, Saeed Vahidian, Hancheng Ye, Jianyang Gu, Jianyi Zhang, Yiran Chen
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 |
|---|---|---|---|
| use similarity-guided instead of stability-guided prediction on zero-shot TruthfulQA (ablation) | CoreInfer (OPT-6.7b) · TruthfulQA zero-shot | BLEU max: 6.22 -> 9.12 | Improved |
| use similarity-guided instead of stability-guided prediction on zero-shot TriviaQA (ablation) | CoreInfer (OPT-6.7b) · TriviaQA zero-shot | exact match: 12.0 -> 21.8 | Improved |
| use similarity-guided instead of stability-guided prediction on zero-shot wmt16-de-en (ablation) | CoreInfer (OPT-6.7b) · wmt16-de-en zero-shot | BLEU: 12.2 -> 26.3 | Improved |
| use similarity-guided instead of stability-guided prediction on zero-shot wmt16-ro-en (ablation) | CoreInfer (OPT-6.7b) · wmt16-ro-en zero-shot | BLEU: 3.36 -> 27.8 | Improved |
| use stability-guided prediction on few-shot and information extraction tasks (ablation) | CoreInfer (OPT-6.7b) · Xsum / SQuAD / few-shot QA and translation | task metrics: 6.7 -> 6.3 Xsum rouge; 52.1 -> 53.2 SQuAD | No clear effect |
| apply CoreInfer sentence-level sparse inference | OPT-6.7b · Xsum | rouge: 6.7 -> 6.3 | No clear effect |
| apply CoreInfer to LLaMA3.1-8b | LLaMA3.1-8b · SQuAD | contains: 54.3 -> 49.7 | Hurt |
| apply CoreInfer to OPT-13b | OPT-13b · wmt16-de-en zero-shot | BLEU: 31.3 -> 35.2 | Improved |
| use CoreInfer instead of PowerInfer on TITAN XP | OPT-6.7b · decode speed | tokens/s: 7.32 -> 19.83 | Improved |
| use CoreInfer instead of Huggingface Transformer on TITAN XP | OPT-6.7b · speedup | speedup: 1x -> 10.33x | Improved |
| increase core neuron ratio for translation (beta = gamma 0.2 to 0.4) (ablation) | CoreInfer · Translation tasks | BLEU: about 40% of neurons | Improved |
Closest papers by meaning
- ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Models (ICLR 2024 · similarity 0.91)
- Learn To be Efficient: Build Structured Sparsity in Large Language Models (NeurIPS 2024 · similarity 0.90)
- R-Sparse: Rank-Aware Activation Sparsity for Efficient LLM Inference (ICLR 2025 · similarity 0.90)
- Sparsing Law: Towards Large Language Models with Greater Activation Sparsity (ICML 2025 · similarity 0.90)
- The Sparsity Whisperer (NeurIPS 2026 · similarity 0.89)
- Pruning Large Language Models with Semi-Structural Adaptive Sparse Training (AAAI 2025 · similarity 0.89)
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.
