© 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. 19002
    Prev
    Next
    INDEX
    Explanations

    Non-English characters

    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
    roph
    -0.26
     wanna
    -0.26
     Morgan
    -0.26
    rcode
    -0.25
    outu
    -0.24
    åĬłéĩį
    -0.24
    áģ
    -0.23
    @return
    -0.23
    ostream
    -0.23
    levard
    -0.23
    POSITIVE LOGITS
    åł°
    0.29
    ç¼Ĩ
    0.29
    工夫
    0.27
    ç¥ļ
    0.27
    itted
    0.27
     Beam
    0.27
    è¾ķ
    0.25
    .axis
    0.24
    åľŁèĢ³
    0.24
    unded
    0.24
    Activations Density 0.051%

    No Known Activations