Pre-computed Q-Filters for efficient KV cache compression.
Nathan Godey
nthngdy
AI & ML interests
None yet
Recent Activity
updated a model 1 day ago
nthngdy/matryoshka-3B updated a model 8 days ago
nthngdy/matryoshka-baselines upvoted a paper 10 days ago
The State-Prediction Separation HypothesisOrganizations
models 63
nthngdy/matryoshka-3B
Text Generation • 3B • Updated • 693
nthngdy/matryoshka-200M
Text Generation • 3B • Updated • 120
nthngdy/copy
Updated
nthngdy/matryoshka-1B
Text Generation • 1B • Updated • 7
nthngdy/matryoshka-baselines
Updated • 29
nthngdy/matritest_1B
Text Generation • 0.6B • Updated • 7
nthngdy/matritest_600M
Text Generation • 0.4B • Updated • 9
nthngdy/matritest_300M
Text Generation • 0.2B • Updated • 10
nthngdy/matritest_van_1B
Text Generation • 1B • Updated • 7
nthngdy/matritest_van_600M
Text Generation • 0.6B • Updated • 8
datasets 30
nthngdy/mmlu_olmo3_contamination
Viewer • Updated • 15.9k • 685
nthngdy/hellaswag_wikihow_olmo_contamination
Viewer • Updated • 1.28k • 12
nthngdy/hellaswag_activitynet
Viewer • Updated • 21.5k • 37
nthngdy/hellaswag_wikihow
Viewer • Updated • 38.4k • 67
nthngdy/mmlu_olmo_contaminated
Viewer • Updated • 3.33k • 627
nthngdy/mmlu_olmo_decontaminated
Viewer • Updated • 10.7k • 508
nthngdy/pile_small_deduped_tokenized
Viewer • Updated • 100k • 47
nthngdy/mmlu_olmo_contamination
Viewer • Updated • 15.9k • 539
nthngdy/mmlu_shuffled
Viewer • Updated • 31.7k • 626
nthngdy/mmlu
Viewer • Updated • 116k • 21