© 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. 25-TRANSCODER-HP
    7. 108706
    Prev
    Next
    INDEX
    Explanations

    say "the number 2"

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_25.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.25.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    idot
    -0.28
    个乡éķĩ
    -0.27
    éĥ¨éĥ¨éķ¿
    -0.26
    cola
    -0.26
     dokÅĤadnie
    -0.26
    å¼Ł
    -0.25
    ades
    -0.24
    oxid
    -0.24
    acher
    -0.24
    çīĪæĿĥå½Ĵ
    -0.23
    POSITIVE LOGITS
     offence
    0.28
    aly
    0.28
    è¾ħ
    0.27
    ymm
    0.26
    æ³Ĭ
    0.26
     Fallon
    0.25
    gence
    0.24
    è¾ħåĬ©
    0.24
     unicode
    0.24
    年人
    0.23
    Activations Density 0.904%

    No Known Activations