© 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. 42957
    Prev
    Next
    INDEX
    Explanations

    snippets of text

    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
    ãĤªãĤ¹ãĤ¹ãĥ¡
    -0.28
     retarded
    -0.27
    溲
    -0.27
    å¢Ļå£ģ
    -0.26
    Sortable
    -0.25
    vincial
    -0.25
    indows
    -0.25
    毵
    -0.24
    Compatible
    -0.24
    volent
    -0.24
    POSITIVE LOGITS
     always
    0.30
     cr
    0.28
     way
    0.28
    æĸ°é²ľ
    0.27
    nung
    0.26
     Myst
    0.26
    always
    0.25
     ABS
    0.24
    stay
    0.24
    æĸ°
    0.24
    Activations Density 0.004%

    No Known Activations