Semantic Density Effect (SDE): Maximizing Information Per Token Improves LLM Accuracy
Fuente:
arXiv
Guardado en:
| Autor principal: | Ahmed, Amr |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
Ejemplares similares
Rethinking Supervised Fine-Tuning: Emphasizing Key Answer Tokens for Improved LLM Accuracy
por: Shi, Xiaofeng, et al.
Publicado: (2025)
por: Shi, Xiaofeng, et al.
Publicado: (2025)
LLM as a Broken Telephone: Iterative Generation Distorts Information
por: Mohamed, Amr, et al.
Publicado: (2025)
por: Mohamed, Amr, et al.
Publicado: (2025)
SP^2DPO: An LLM-assisted Semantic Per-Pair DPO Generalization
por: He, Chaoyue, et al.
Publicado: (2026)
por: He, Chaoyue, et al.
Publicado: (2026)
MoLoRA: Composable Specialization via Per-Token Adapter Routing
por: Shah, Shrey, et al.
Publicado: (2026)
por: Shah, Shrey, et al.
Publicado: (2026)
Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining
por: Feng, Steven, et al.
Publicado: (2024)
por: Feng, Steven, et al.
Publicado: (2024)
Maximizing Mutual Information Between Prompt and Response Improves LLM Performance With No Additional Data
por: Nam, Hyunji, et al.
Publicado: (2026)
por: Nam, Hyunji, et al.
Publicado: (2026)
MIG: Automatic Data Selection for Instruction Tuning by Maximizing Information Gain in Semantic Space
por: Chen, Yicheng, et al.
Publicado: (2025)
por: Chen, Yicheng, et al.
Publicado: (2025)
Demystifying Reasoning Dynamics with Mutual Information: Thinking Tokens are Information Peaks in LLM Reasoning
por: Qian, Chen, et al.
Publicado: (2025)
por: Qian, Chen, et al.
Publicado: (2025)
Broken-Token: Filtering Obfuscated Prompts by Counting Characters-Per-Token
por: Zychlinski, Shaked, et al.
Publicado: (2025)
por: Zychlinski, Shaked, et al.
Publicado: (2025)
Semantic Tokens in Retrieval Augmented Generation
por: Suro, Joel
Publicado: (2024)
por: Suro, Joel
Publicado: (2024)
Revisiting the Uniform Information Density Hypothesis in LLM Reasoning
por: Gwak, Minju, et al.
Publicado: (2025)
por: Gwak, Minju, et al.
Publicado: (2025)
SemToken: Semantic-Aware Tokenization for Efficient Long-Context Language Modeling
por: Liu, Dong, et al.
Publicado: (2025)
por: Liu, Dong, et al.
Publicado: (2025)
CATP: Cross-Attention Token Pruning for Accuracy Preserved Multimodal Model Inference
por: Liao, Ruqi, et al.
Publicado: (2024)
por: Liao, Ruqi, et al.
Publicado: (2024)
The Silent Vote: Improving Zero-Shot LLM Reliability by Aggregating Semantic Neighborhoods
por: Badhe, Sanket, et al.
Publicado: (2026)
por: Badhe, Sanket, et al.
Publicado: (2026)
Know When To Fold 'Em: Token-Efficient LLM Synthetic Data Generation via Multi-Stage In-Flight Rejection
por: Chowdhury, Anjir Ahmed, et al.
Publicado: (2026)
por: Chowdhury, Anjir Ahmed, et al.
Publicado: (2026)
When Safety Blocks Sense: Measuring Semantic Confusion in LLM Refusals
por: Anonto, Riad Ahmed, et al.
Publicado: (2025)
por: Anonto, Riad Ahmed, et al.
Publicado: (2025)
PersLLM: A Personified Training Approach for Large Language Models
por: Zeng, Zheni, et al.
Publicado: (2024)
por: Zeng, Zheni, et al.
Publicado: (2024)
Extending Token Computation for LLM Reasoning
por: Liao, Bingli, et al.
Publicado: (2024)
por: Liao, Bingli, et al.
Publicado: (2024)
Discovering Multi-Scale Semantic Structure in Text Corpora Using Density-Based Trees and LLM Embeddings
por: Haschka, Thomas, et al.
Publicado: (2025)
por: Haschka, Thomas, et al.
Publicado: (2025)
One Token Is Enough: Improving Diffusion Language Models with a Sink Token
por: Zhang, Zihou, et al.
Publicado: (2026)
por: Zhang, Zihou, et al.
Publicado: (2026)
TimeStampEval: A Simple LLM Eval and a Little Fuzzy Matching Trick to Improve Search Accuracy
por: McCammon, James
Publicado: (2025)
por: McCammon, James
Publicado: (2025)
InFact: Informativeness Alignment for Improved LLM Factuality
por: Cohen, Roi, et al.
Publicado: (2025)
por: Cohen, Roi, et al.
Publicado: (2025)
AI-Augmented Predictions: LLM Assistants Improve Human Forecasting Accuracy
por: Schoenegger, Philipp, et al.
Publicado: (2024)
por: Schoenegger, Philipp, et al.
Publicado: (2024)
KV Prediction for Improved Time to First Token
por: Horton, Maxwell, et al.
Publicado: (2024)
por: Horton, Maxwell, et al.
Publicado: (2024)
LLM-Oriented Token-Adaptive Knowledge Distillation
por: Xie, Xurong, et al.
Publicado: (2025)
por: Xie, Xurong, et al.
Publicado: (2025)
HeartLLM: Discretized ECG Tokenization for LLM-Based Diagnostic Reasoning
por: Yang, Jinning, et al.
Publicado: (2025)
por: Yang, Jinning, et al.
Publicado: (2025)
Improving Speech Recognition of Named Entities in Classroom Speech with LLM Revision and Phonetic-Semantic Context
por: Trinh, Viet Anh, et al.
Publicado: (2025)
por: Trinh, Viet Anh, et al.
Publicado: (2025)
Think-Augmented Function Calling: Improving LLM Parameter Accuracy Through Embedded Reasoning
por: Wei, Lei, et al.
Publicado: (2026)
por: Wei, Lei, et al.
Publicado: (2026)
SePer: Measure Retrieval Utility Through The Lens Of Semantic Perplexity Reduction
por: Dai, Lu, et al.
Publicado: (2025)
por: Dai, Lu, et al.
Publicado: (2025)
Token Masking Improves Transformer-Based Text Classification
por: Xu, Xianglong, et al.
Publicado: (2025)
por: Xu, Xianglong, et al.
Publicado: (2025)
SD-E$^2$: Semantic Exploration for Reasoning Under Token Budgets
por: Mishra, Kshitij, et al.
Publicado: (2026)
por: Mishra, Kshitij, et al.
Publicado: (2026)
Explainability-Based Token Replacement on LLM-Generated Text
por: Mohammadi, Hadi, et al.
Publicado: (2025)
por: Mohammadi, Hadi, et al.
Publicado: (2025)
Enhancing Large Language Models for Mobility Analytics with Semantic Location Tokenization
por: Chen, Yile, et al.
Publicado: (2025)
por: Chen, Yile, et al.
Publicado: (2025)
GPS: General Per-Sample Prompter
por: Batorski, Pawel, et al.
Publicado: (2025)
por: Batorski, Pawel, et al.
Publicado: (2025)
AlphaToken: Decoupling Adaptation and Stability for Path-Aware Response Token Valuation in LLM Post-Training
por: Qing, Liu, et al.
Publicado: (2026)
por: Qing, Liu, et al.
Publicado: (2026)
Learning to Maximize Mutual Information for Chain-of-Thought Distillation
por: Chen, Xin, et al.
Publicado: (2024)
por: Chen, Xin, et al.
Publicado: (2024)
Beyond Accuracy: The Role of Calibration in Self-Improving Large Language Models
por: Huang, Liangjie, et al.
Publicado: (2025)
por: Huang, Liangjie, et al.
Publicado: (2025)
Reinforcement Learning vs. Distillation: Understanding Accuracy and Capability in LLM Reasoning
por: Kim, Minwu, et al.
Publicado: (2025)
por: Kim, Minwu, et al.
Publicado: (2025)
Blended RAG: Improving RAG (Retriever-Augmented Generation) Accuracy with Semantic Search and Hybrid Query-Based Retrievers
por: Sawarkar, Kunal, et al.
Publicado: (2024)
por: Sawarkar, Kunal, et al.
Publicado: (2024)
Tokenization Matters: Improving Zero-Shot NER for Indic Languages
por: Pattnayak, Priyaranjan, et al.
Publicado: (2025)
por: Pattnayak, Priyaranjan, et al.
Publicado: (2025)
Ejemplares similares
-
Rethinking Supervised Fine-Tuning: Emphasizing Key Answer Tokens for Improved LLM Accuracy
por: Shi, Xiaofeng, et al.
Publicado: (2025) -
LLM as a Broken Telephone: Iterative Generation Distorts Information
por: Mohamed, Amr, et al.
Publicado: (2025) -
SP^2DPO: An LLM-assisted Semantic Per-Pair DPO Generalization
por: He, Chaoyue, et al.
Publicado: (2026) -
MoLoRA: Composable Specialization via Per-Token Adapter Routing
por: Shah, Shrey, et al.
Publicado: (2026) -
Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining
por: Feng, Steven, et al.
Publicado: (2024)