© 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. 19261
    Prev
    Next
    INDEX
    Explanations

    .

    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
    æĶ¾è¿ĩ
    -0.26
    pus
    -0.26
    .Inject
    -0.24
    IAN
    -0.24
    á»ijn
    -0.23
    antan
    -0.23
    borg
    -0.23
    æģķ
    -0.22
     repaint
    -0.22
    åıŁ
    -0.22
    POSITIVE LOGITS
    th
    0.32
     mm
    0.30
    äºĮåįģå¹´
    0.30
    mm
    0.30
    毫米
    0.26
    mens
    0.26
    æŁ¥æī¾
    0.26
    亥
    0.25
    æĥĬ
    0.25
    VarChar
    0.25
    Activations Density 0.001%

    No Known Activations