Picture for Feiyang Kang

Feiyang Kang

AutoScale: Automatic Prediction of Compute-optimal Data Composition for Training LLMs

Add code
Jul 29, 2024
Viaarxiv icon

Get more for less: Principled Data Selection for Warming Up Fine-Tuning in LLMs

Add code
May 05, 2024
Viaarxiv icon

FASTTRACK: Fast and Accurate Fact Tracing for LLMs

Add code
Apr 22, 2024
Viaarxiv icon

The Mirrored Influence Hypothesis: Efficient Data Influence Estimation by Harnessing Forward Passes

Add code
Feb 14, 2024
Viaarxiv icon

Data Acquisition: A New Frontier in Data-centric AI

Add code
Nov 22, 2023
Viaarxiv icon

Performance Scaling via Optimal Transport: Enabling Data Selection from Partially Revealed Sources

Add code
Jul 05, 2023
Viaarxiv icon

LAVA: Data Valuation without Pre-Specified Learning Algorithms

Add code
Apr 28, 2023
Figure 1 for LAVA: Data Valuation without Pre-Specified Learning Algorithms
Figure 2 for LAVA: Data Valuation without Pre-Specified Learning Algorithms
Figure 3 for LAVA: Data Valuation without Pre-Specified Learning Algorithms
Figure 4 for LAVA: Data Valuation without Pre-Specified Learning Algorithms
Viaarxiv icon