Distilled models from paper Thinking at the Right Size: Amortized Distillation Across Post-Trained LLMs (https://arxiv.org/abs/2608.22854).
AI & ML interests
Data-Centric ML
Recent Activity
View all activity
Papers
A Critical Look at Targeted Instruction Selection: Disentangling What Matters (and What Doesn't)
Boomerang Distillation Enables Zero-Shot Model Size Interpolation
models 11
Harvard-DCML/ADAPT-Qwen3-8.5B
Text Generation • 0.5B • Updated • 22
Harvard-DCML/ADAPT-Qwen3-8.5B-Base
Text Generation • 0.5B • Updated • 16
Harvard-DCML/ADAPT-Llama3.1-4.8B-Instruct
Text Generation • 1B • Updated • 12
Harvard-DCML/ADAPT-Olmo3-4.3B-Instruct
Text Generation • 1B • Updated • 13
Harvard-DCML/ADAPT-Qwen3-2.3B-Instruct
Text Generation • 0.6B • Updated • 17
Harvard-DCML/ADAPT-Qwen3-2.3B-Base
Text Generation • 0.6B • Updated • 30
Harvard-DCML/boomerang-qwen3-2.3B
Text Generation • 3B • Updated • 738 • 3
Harvard-DCML/boomerang-qwen3-4.9B
5B • Updated • 18
Harvard-DCML/boomerang-pythia-3.8B
4B • Updated • 11
Harvard-DCML/boomerang-pythia-1.6B
2B • Updated • 19
datasets 31
Harvard-DCML/tis-dolci-random-unbalanced
Viewer • Updated • 30k • 24
Harvard-DCML/tis-dolci-subset-datasets-gtr-t5-base
Viewer • Updated • 50k • 17
Harvard-DCML/tis-dolci-subset-datasets-Olmo-3-1025-7B
Viewer • Updated • 300k • 102
Harvard-DCML/tis-dolci-subset-datasets-SmolLM3-3B-Base
Viewer • Updated • 300k • 100
Harvard-DCML/tis-dolci-subset-datasets-Qwen3-4B-Base
Viewer • Updated • 300k • 249
Harvard-DCML/tis-dolci-subset-datasets-Llama-3.2-3B
Viewer • Updated • 300k • 16
Harvard-DCML/tis-dolci-subset-datasets-Llama-2-7b-hf
Viewer • Updated • 300k • 34
Harvard-DCML/tis-dolci-quantile-datasets-gtr-t5-base
Updated • 61
Harvard-DCML/tis-dolci-quantile-datasets-Olmo-3-1025-7B
Updated • 39
Harvard-DCML/tis-dolci-quantile-datasets-SmolLM3-3B-Base
Updated • 8