juiceb0xc0de/juicebucket
0 Bytes
R PRO
juiceb0xc0de
AI & ML interests
You can call me McDreamy M.D. I'll be your attending neural network brain surgeon.
Recent Activity
updated a model 39 minutes ago
juiceb0xc0de/smollm2-135m-SAE published a model 40 minutes ago
juiceb0xc0de/smollm2-135m-SAE updated a dataset about 1 hour ago
juiceb0xc0de/smollm2-135m-pretokOrganizations
Posts 13
Post
109
I've been busy testing the event aware SAE trainer I've been designing. It uses hierarchical supervisory control for hands-off SAE training taking the pressure of having to tune hyperparameters per layer away. The trainer uses an Adaptive Lagrangian controller to drive L0 sparsity to target automatically, with Explained Variance as the convergence quality metric.
Last night I successfully trained SAE weights for all 30 layers of HuggingFaceTB/SmolLM2-135M-Instruct without adjusting a single hyperparameter. This was all automated, required no supervision.
Using an A100-40GB GPU I maintained ~400 tok/s throughput and was finished in 2 hours and 33 minutes! I ended the run off with 0 dead features across all layers. L0 convergence was set at 50 and mean Explained Variance of 0.9519 with the ceiling converging at 0.9905!
SAE weights and full model SQLite database live now!
juiceb0xc0de/smollm2-135m-instruct-SAE (weights only)
juiceb0xc0de/smollm2-135m-instruct-atlas-SAE (weights plus custom atlas + SQLite database)
Last night I successfully trained SAE weights for all 30 layers of HuggingFaceTB/SmolLM2-135M-Instruct without adjusting a single hyperparameter. This was all automated, required no supervision.
Using an A100-40GB GPU I maintained ~400 tok/s throughput and was finished in 2 hours and 33 minutes! I ended the run off with 0 dead features across all layers. L0 convergence was set at 50 and mean Explained Variance of 0.9519 with the ceiling converging at 0.9905!
SAE weights and full model SQLite database live now!
juiceb0xc0de/smollm2-135m-instruct-SAE (weights only)
juiceb0xc0de/smollm2-135m-instruct-atlas-SAE (weights plus custom atlas + SQLite database)
This is a collection of models that I've trained on data collected through conversations with frontier models GPT, Claude, Perplexity and myself.
-
juiceb0xc0de/locus-gemma-4-e2b
Text Generation • 8B • Updated • 24 • 3 -
juiceb0xc0de/bella-bartender-gemma-e2b
Text Generation • 5B • Updated • 72 • 4 -
juiceb0xc0de/bella-bartender-8b-llama3.1
Text Generation • 8B • Updated • 170 • • 4 -
juiceb0xc0de/bella-bartender-heretic-1b
Text Generation • 1B • Updated • 248 • 2
De-tuned models
This is a collection of models that I've trained on data collected through conversations with frontier models GPT, Claude, Perplexity and myself.
-
juiceb0xc0de/locus-gemma-4-e2b
Text Generation • 8B • Updated • 24 • 3 -
juiceb0xc0de/bella-bartender-gemma-e2b
Text Generation • 5B • Updated • 72 • 4 -
juiceb0xc0de/bella-bartender-8b-llama3.1
Text Generation • 8B • Updated • 170 • • 4 -
juiceb0xc0de/bella-bartender-heretic-1b
Text Generation • 1B • Updated • 248 • 2
spaces 4
pinned
Sleeping
Agents
2
LR Scheduler Benchmark
🏁
Standardized benchmark comparing learning rate schedulers
Sleeping
Agents
7
Gemma 4 E2b Brain Atlas
👀
Serve a Gradio web UI with embedded HF token
Sleeping
2
Forge
🔥
Build compatible ML training recipes with evidence
Sleeping
Agents
Trackio
🚀
Track and visualize data sequences with interactive displays
models 18
juiceb0xc0de/smollm2-135m-SAE
Updated
juiceb0xc0de/bella-bartender-gemma-e4b-GGUF
Text Generation • 7B • Updated • 582 • 1
juiceb0xc0de/locus-gemma-4-e2b
Text Generation • 8B • Updated • 24 • 3
juiceb0xc0de/bella-bartender-gemma-e4b
8B • Updated • 8 • 1
juiceb0xc0de/bella-e4b-subzero-v1
Text Generation • 8B • Updated • 1
juiceb0xc0de/bella-bartender-gemma-e2b
Text Generation • 5B • Updated • 72 • 4
juiceb0xc0de/bella-bartender-gemma-4-e2b-gguf
Text Generation • 5B • Updated • 201 • 1
juiceb0xc0de/bella-bartender-8b-llama3.1
Text Generation • 8B • Updated • 170 • • 4
juiceb0xc0de/bella-bartender-3b
Text Generation • 3B • Updated • 246 • 2
juiceb0xc0de/bella-bartender-v2-8b
Text Generation • 8B • Updated • 84 • 4
datasets 11
juiceb0xc0de/smollm2-135m-pretok
Updated
juiceb0xc0de/smollm2-135m-instruct-SAE
Viewer • Updated • 30 • 24
juiceb0xc0de/smollm2-135m-instruct-atlas-SAE
Updated • 869 • 1
juiceb0xc0de/qwen3-8b-base-atlas-SAE
Updated • 1.42k • 1
juiceb0xc0de/mapping-prompts
Preview • Updated • 85
juiceb0xc0de/qwen3-8b-atlas
Updated • 1.54k
juiceb0xc0de/gemma-4-e2b-saes
Viewer • Updated • 9 • 141
juiceb0xc0de/gemma-4-e2b-saes-rolling
Viewer • Updated • 4 • 94
juiceb0xc0de/gemma-4-e2b-atlas
Viewer • Updated • 1.3M • 184 • 4
juiceb0xc0de/chaotic-absurdity
Viewer • Updated • 100 • 28