Adaptive Computation Modules: Granular Conditional Computation For Efficient Inference
Paper
•
2312.10193
•
Published
•
1
Alessio
alessiodevoto
AI & ML interests
Efficient training/inference - Adaptive computation - Interpretability
Organizations
authored
4
papers
9 months ago
A Simple and Effective $L_2$ Norm-Based Strategy for KV Cache Compression
Paper
•
2406.11430
•
Published
•
24
Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering
Paper
•
2410.15999
•
Published
•
20
Analysing the Residual Stream of Language Models Under Knowledge Conflicts
Paper
•
2410.16090
•
Published
•
7
authored
a
paper
about 1 year ago