Picture for Dan Alistarh

Dan Alistarh

Efficient Data Selection at Scale via Influence Distillation

Add code
May 25, 2025
Viaarxiv icon

SVD-Free Low-Rank Adaptive Gradient Optimization for Large Language Models

Add code
May 23, 2025
Viaarxiv icon

Layer-wise Quantization for Quantized Optimistic Dual Averaging

Add code
May 20, 2025
Viaarxiv icon

Quartet: Native FP4 Training Can Be Optimal for Large Language Models

Add code
May 20, 2025
Viaarxiv icon

Towards Combinatorial Interpretability of Neural Computation

Add code
Apr 10, 2025
Viaarxiv icon

Hogwild! Inference: Parallel LLM Generation via Concurrent Attention

Add code
Apr 09, 2025
Viaarxiv icon

Compression Scaling Laws:Unifying Sparsity and Quantization

Add code
Feb 23, 2025
Viaarxiv icon

DarwinLM: Evolutionary Structured Pruning of Large Language Models

Add code
Feb 11, 2025
Viaarxiv icon

Position: It's Time to Act on the Risk of Efficient Personalized Text Generation

Add code
Feb 10, 2025
Viaarxiv icon

QuEST: Stable Training of LLMs with 1-Bit Weights and Activations

Add code
Feb 07, 2025
Figure 1 for QuEST: Stable Training of LLMs with 1-Bit Weights and Activations
Figure 2 for QuEST: Stable Training of LLMs with 1-Bit Weights and Activations
Figure 3 for QuEST: Stable Training of LLMs with 1-Bit Weights and Activations
Figure 4 for QuEST: Stable Training of LLMs with 1-Bit Weights and Activations
Viaarxiv icon