© 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. 96707
    Prev
    Next
    INDEX
    Explanations

    describing something

    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
    å¼Ģåĩº
    -0.28
     READY
    -0.27
    ä¾Ľ
    -0.27
    åĬłæ²¹
    -0.27
    ç¼ĸåĨĻ
    -0.27
    _hot
    -0.26
    ibilit
    -0.25
     Princip
    -0.25
     feasibility
    -0.25
    comed
    -0.25
    POSITIVE LOGITS
     possesses
    0.35
     retains
    0.35
     was
    0.32
     includes
    0.32
     incorporates
    0.31
     retain
    0.31
     preserves
    0.31
     contains
    0.31
    $ret
    0.31
     consists
    0.30
    Activations Density 0.908%

    No Known Activations