© 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. 96244
    Prev
    Next
    INDEX
    Explanations

    Due

    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
     tuning
    -0.27
    é¦ĸå±Ĭ
    -0.25
     LU
    -0.25
    crm
    -0.25
     tun
    -0.24
    handled
    -0.24
     cases
    -0.23
    äºĭ迹
    -0.23
    üns
    -0.23
    è°¤
    -0.23
    POSITIVE LOGITS
    注æĦı
    0.32
    西åįĹ
    0.29
     instead
    0.28
    out
    0.28
    西
    0.28
    Instead
    0.27
     Instead
    0.27
    åħ³
    0.27
    åĩ¡
    0.27
    berman
    0.26
    Activations Density 0.890%

    No Known Activations