Single-seed pilot comparing frozen LLM-judge and probe monitors. Checkpoints and evidence; no established safety guarantee.
Burny
BurnyCoder
AI & ML interests
deep learning, LLMs, interpretability, science, physics
Recent Activity
liked a model about 3 hours ago
RedHatAI/GLM-5.3-MXFP4xFP8 liked a model about 5 hours ago
inclusionAI/Ling-3.0-tiny-UBTree updated a collection about 6 hours ago
Qwen3-0.6B GRAFT with LLM-as-a-Judge and Probe MonitorsOrganizations
models 37
BurnyCoder/graft-qwen3-06b-random-graft-pilot
Updated
BurnyCoder/graft-qwen3-06b-probe-filtering-pilot
Updated
BurnyCoder/graft-qwen3-06b-original-graft-pilot
Updated
BurnyCoder/graft-qwen3-06b-qwen-judge-graft-pilot
Updated • 1
BurnyCoder/graft-qwen3-06b-probe-graft-pilot
Updated • 1
BurnyCoder/graft-qwen3-06b-no-intervention-pilot
Updated • 1
BurnyCoder/grokking-modular-addition-runs
Updated
BurnyCoder/oscillatory-flow-matching-mnist
Unconditional Image Generation • Updated
BurnyCoder/oscillatory-flow-matching-cifar10
Unconditional Image Generation • Updated
BurnyCoder/rl-locomotion-steering-vectors
Reinforcement Learning • Updated