© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 31-TRANSCODER-HP
    7. 19587
    Prev
    Next
    INDEX
    Explanations

    un

    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
    æŀī
    -0.26
    ignKey
    -0.26
    hek
    -0.26
    icable
    -0.24
    >Note
    -0.24
     apt
    -0.24
    éĩĩ
    -0.24
    ish
    -0.24
    fast
    -0.23
    ifty
    -0.23
    POSITIVE LOGITS
    æŃ¢
    0.29
     constitu
    0.27
    bond
    0.25
    追éĢIJ
    0.25
    à¸ĵะ
    0.25
    åľ¨åĮĹ京
    0.24
    /browser
    0.24
    :].
    0.24
    éĻIJåĪ¶
    0.24
    achs
    0.24
    Activations Density 0.142%

    No Known Activations