© 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. 32-TRANSCODER-HP
    7. 43720
    Prev
    Next
    INDEX
    Explanations

    1

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     compositions
    -0.32
     composition
    -0.28
     Composition
    -0.27
    iscal
    -0.27
    ãĤ¨ãĥ«
    -0.27
    á»Ħ
    -0.26
     mind
    -0.25
    oc
    -0.25
    èģĨ
    -0.25
    El
    -0.25
    POSITIVE LOGITS
     Dix
    0.28
     Pit
    0.28
     Trent
    0.28
     tph
    0.27
    kre
    0.27
    è¾¾
    0.26
    avr
    0.26
    åħ¸
    0.26
    (EFFECT
    0.26
    akh
    0.25
    Activations Density 0.005%

    No Known Activations