Sara Kangaslahti's picture

Sara Kangaslahti

s-kangasl

·

AI & ML interests

None yet

Recent Activity

liked a model 19 days ago

Harvard-DCML/boomerang-qwen3-2.3B

new activity 5 months ago

Harvard-DCML/boomerang-qwen3-4.9B:Substantially lower accuracy on reasoning benchmarks such as GSM8K (1.5%) and MATH-500 (4.2%)

authored a paper 7 months ago

Can You Label Less by Using Out-of-Domain Data? Active & Transfer Learning with Few-shot Instructions

View all activity

Organizations

upvoted a collection 7 months ago

Boomerang Distillation

Distilled models from the boomerang distillation paper (https://arxiv.org/abs/2510.05064). • 6 items • Updated Oct 14, 2025 • 1

upvoted a paper 7 months ago

Boomerang Distillation Enables Zero-Shot Model Size Interpolation

Paper • 2510.05064 • Published Oct 6, 2025 • 1