© 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. 31-TRANSCODER-HP
    7. 19437
    Prev
    Next
    INDEX
    Explanations

    scientific names / identifiers

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    loor
    -0.30
    ãĥ¡ãĥ³ãĥĪ
    -0.30
    æıĴ
    -0.28
    ÑĮеÑĢ
    -0.26
    .stem
    -0.26
    _plate
    -0.26
    plate
    -0.26
    ignite
    -0.26
    itioner
    -0.26
    plug
    -0.26
    POSITIVE LOGITS
     cơ
    0.27
    edb
    0.27
    oda
    0.26
    emie
    0.26
    åģļäºĭ
    0.25
    rif
    0.25
     Eden
    0.25
     crit
    0.24
     grown
    0.24
     am
    0.24
    Activations Density 0.011%

    No Known Activations