Distilling Sequential Computation in Transformer Language Models

arXiv:2609.27233 · cs.CL · Submitted 2026-09-23 · Read on arXiv

cs.CL

Submitted: 2026-09-23

Updated: 2026-09-23

Code: https://github.com/Zesearch/Umim-LLM

Project page: http://skylion007.github.io/OpenWebTextCorpus

Terminology

• Embeddings — 85× in this paper · explained in 2 episodes
• Adaptation — 62× in this paper · explained in 1 episode
• Distillation — 53× in this paper · explained in 2 episodes
• Transformer — 44× in this paper · explained in 2 episodes
• Accuracy — 40× in this paper · explained in 1 episode
• Alignment — 24× in this paper · explained in 1 episode
• Ablation — 20× in this paper · explained in 1 episode
• KV cache — 19× in this paper · explained in 3 episodes
• Fine-Tuning — 13× in this paper · explained in 5 episodes
• Downstream Tasks — 12× in this paper · explained in 2 episodes
• Tokenization — 11× in this paper · explained in 2 episodes
• Generalization — 10× in this paper · explained in 13 episodes
• Perplexity — 8× in this paper · explained in 3 episodes
• Decode — 5× in this paper · explained in 2 episodes
• Pruning — 5× in this paper · explained in 4 episodes
• Benchmark — 4× in this paper · explained in 2 episodes
• Cosine Similarity — 4× in this paper · explained in 3 episodes
• Frozen model — 4× in this paper · explained in 3 episodes
• Ablation study — 3× in this paper · explained in 1 episode
• Adam — 3× in this paper · explained in 2 episodes
• Distribution Shift — 3× in this paper · explained in 5 episodes
• Semantic Similarity — 3× in this paper · explained in 2 episodes
• Mean Reciprocal Rank — 2× in this paper
• Regularization — 2× in this paper · explained in 3 episodes
• Transformer Architecture — 2× in this paper · explained in 2 episodes
• Absolute Indices — 1× in this paper · explained in 1 episode
• Alignment Scores — 1× in this paper · explained in 1 episode
• Byte Pair Encoding — 1× in this paper
• Causal Reasoning — 1× in this paper · explained in 2 episodes
• Clustering — 1× in this paper · explained in 2 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Domain Adaptation — 1× in this paper · explained in 5 episodes
• Domain Shift — 1× in this paper · explained in 4 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Model-Agnostic — 1× in this paper · explained in 2 episodes
• Parameter Efficiency — 1× in this paper · explained in 3 episodes
• Parameter-Efficient Fine-Tuning — 1× in this paper · explained in 2 episodes
• Reasoning Traces — 1× in this paper · explained in 3 episodes
• Sparsity — 1× in this paper · explained in 4 episodes
• Trace — 1× in this paper · explained in 2 episodes

Sources

• Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
• Efficient Prompt Compression with Evaluator Heads for Long-Context Transformer Inference
• ROUGE 2.0: Updated and Improved Measures for Evaluation of Summarization Tasks
• zip2zip: Inference-Time Adaptive Tokenization via Online Compression
• The Llama 3 Herd of Models
• LLMLingua: Compressing Prompts for Accelerated Inference of Large Language Models
• MrT5: Dynamic Token Merging for Efficient Byte-level Language Models
• Copy Is All You Need
• Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference
• Unlocking Context Constraints of LLMs: Enhancing Context Efficiency of LLMs with Self-Information-Based Content Filtering
• Compressing Context to Enhance Inference Efficiency of Large Language Models
• 500xCompressor: Generalized Prompt Compression for Large Language Models
• Prompt Compression with Context-Aware Sentence Encoding for Fast and Improved LLM Inference
• SuperBPE: Space Travel for Language Models
• Pointer Sentinel Mixture Models
• Byte Latent Transformer: Patches Scale Better Than Tokens
• LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression
• Flexibly Scaling Large Language Models Contexts Through Extensible Tokenization
• ProCut: LLM Prompt Compression via Attribution Estimation
• Cmprsr: Abstractive Token-Level Question-Agnostic Prompt Compressor

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering