© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 30-TRANSCODER-HP
    7. 96279
    Prev
    Next
    INDEX
    Explanations

    Mish

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    TER
    -0.29
    èĪĮå°ĸ
    -0.26
    æľį
    -0.25
     gaz
    -0.24
    atte
    -0.24
    amo
    -0.24
    -factor
    -0.24
    åĽłç´ł
    -0.24
    agt
    -0.24
    illet
    -0.24
    POSITIVE LOGITS
     royalty
    0.26
     care
    0.25
     props
    0.25
    é¡·
    0.25
    å¤Ħå¤Ħ
    0.24
     apart
    0.24
    ::$_
    0.24
     Ple
    0.24
    ().__
    0.23
     breasts
    0.23
    Activations Density 0.004%

    No Known Activations