© 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. 96089
    Prev
    Next
    INDEX
    Explanations

    technical texts

    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
     grop
    -0.27
    Aff
    -0.26
    èĥ¶
    -0.26
    ören
    -0.25
    ç½
    -0.25
    oundation
    -0.24
    _periods
    -0.24
    åıįåĵį
    -0.24
     glue
    -0.24
    ymb
    -0.24
    POSITIVE LOGITS
    æīģ
    0.30
     flat
    0.29
    lice
    0.28
     Heading
    0.27
    heading
    0.26
    flat
    0.26
    åĪĩæį¢
    0.26
    eve
    0.26
     unsupported
    0.25
    ror
    0.25
    Activations Density 0.884%

    No Known Activations