Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
ceselder
/
maxact-fast
like
0
interpretability
sae
activation-steering
max-activating-examples
License:
apache-2.0
Model card
Files
Files and versions
xet
Community
Copy to bucket
new
main
maxact-fast
/
src
/
mxf
13.6 kB
Ctrl+K
Ctrl+K
2 contributors
History:
7 commits
ceselder
pretrain: sequence packing (block-diag attn, per-marker injection) — no-leak + inject proofs pass; MFU bench pending
e6dc020
8 days ago
__init__.py
Safe
0 Bytes
maxact-fast: clean repo — DESIGN + pretraining pipeline (embed/cluster/probe-cache/build-data/pretrain), Dr. GRPO RL to follow
8 days ago
config.py
3.76 kB
rl: true per-token entropy bonus (maximize r + beta*H, explicit diversity knob, no KL)
8 days ago
inject.py
3.37 kB
pretrain: sequence packing (block-diag attn, per-marker injection) — no-leak + inject proofs pass; MFU bench pending
8 days ago
mfu.py
1.16 kB
pretrain: MFU meter + length-bucketing + big-batch + optional compile
8 days ago
prompts.py
2.11 kB
prompts: inoculation-style framing (scoped interpretability-tool purpose)
8 days ago
sae.py
3.2 kB
sae: loader + build_sae_data + eval_sae (cross-uplift metric); weights_only=False for torch 2.13
8 days ago
