Llama-3.1

Llama-3.1 is a benchmark dataset used to evaluate the performance of large language models and vision-language models, particularly in the context of their computational efficiency and accuracy during the prefill and decode stages of inference.

Papers using Llama-3.1 (8)

Information-Aware KV Cache Compression for Long Reasoning2026Geometry-Aware Online Scheduling for LLM Serving: From Theoretical Bound to System Practice2026Improving Quantized Model Performance in Qualitative Analysis with Multi-Pass Prompt Verification2026SCAN: Sparse Circuit Anchor Interpretable Neuron for Lifelong Knowledge Editing2026POP: Prefill-Only Pruning for Efficient Large Model Inference2026Garbage Attention in Large Language Models: BOS Sink Heads and Sink-aware Pruning2026Decomposing MLP Activations into Interpretable Features via Semi-Nonnegative Matrix Factorization2025Precise In-Parameter Concept Erasure in Large Language Models2025