© 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. 42017
    Prev
    Next
    INDEX
    Explanations

    common

    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
    çī¹å®ļ
    -0.27
    ornings
    -0.26
    å¾Īæ·±
    -0.26
     very
    -0.26
    åī§çĥĪ
    -0.26
    å¾Ī
    -0.25
    .original
    -0.25
    å¾Īéķ¿
    -0.24
    pecific
    -0.24
     quite
    -0.24
    POSITIVE LOGITS
     sense
    0.56
    sense
    0.47
    ly
    0.47
    Sense
    0.42
    /common
    0.42
     Sense
    0.41
     senses
    0.39
     common
    0.38
    ally
    0.38
    common
    0.36
    Activations Density 0.020%

    No Known Activations