© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 33-TRANSCODER-HP
    7. 54611
    Prev
    Next
    INDEX
    Explanations

    underscores

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    ^{
    -0.43
    _{
    -0.34
    }_{
    -0.33
     weeds
    -0.27
     rede
    -0.27
     #{
    -0.26
    幸
    -0.26
    è¾¹
    -0.26
    ä¾Ľåħ»
    -0.24
     jump
    -0.24
    POSITIVE LOGITS
    _
    0.65
    __
    0.64
    _',
    0.63
    _",
    0.61
    _=
    0.61
    _.
    0.60
    _,
    0.59
    _↵↵
    0.57
    _↵
    0.56
    _='
    0.56
    Activations Density 0.064%

    No Known Activations