© 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. 19152
    Prev
    Next
    INDEX
    Explanations

    of

    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
    út
    -0.30
    -lat
    -0.30
    ialized
    -0.29
    paces
    -0.27
    TEE
    -0.27
    Normalization
    -0.26
    ñana
    -0.26
    rement
    -0.26
    illac
    -0.26
    .createTextNode
    -0.25
    POSITIVE LOGITS
    val
    0.35
    iko
    0.32
    ãģŀ
    0.28
     atm
    0.27
    èµ·åºĬ
    0.26
    çĽijçĿ£ç®¡çIJĨ
    0.26
     Pra
    0.26
    .val
    0.26
    éĻ¸
    0.26
    mel
    0.26
    Activations Density 1.051%

    No Known Activations