© 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. 19005
    Prev
    Next
    INDEX
    Explanations

    scientific texts

    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
     sensitive
    -0.29
    roller
    -0.28
     sensitivity
    -0.28
    æ´Ĵ
    -0.27
    æķıæĦŁ
    -0.26
    ieve
    -0.25
     Targets
    -0.25
    enza
    -0.25
     consistency
    -0.25
    rollers
    -0.25
    POSITIVE LOGITS
    æĦı
    0.31
    ocos
    0.27
    iku
    0.26
    yp
    0.26
    ìłģìĿ´
    0.26
    yth
    0.26
    èĪ·
    0.25
    .mdl
    0.25
    oris
    0.24
    æĦıæĢĿ
    0.24
    Activations Density 0.041%

    No Known Activations