© 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. 41784
    Prev
    Next
    INDEX
    Explanations

    technical reports

    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
     av
    -0.28
    举
    -0.28
    .av
    -0.26
    å¯Ħ
    -0.26
    aths
    -0.25
    WithPath
    -0.25
    iens
    -0.25
    æ²»
    -0.25
    .Offset
    -0.25
    pan
    -0.24
    POSITIVE LOGITS
     Moor
    0.28
    TON
    0.27
    Ton
    0.26
    缶
    0.25
     Marl
    0.25
    dee
    0.25
    ÄĻd
    0.24
    sst
    0.24
    åħħåĪĨåıijæĮ¥
    0.24
    tons
    0.24
    Activations Density 0.002%

    No Known Activations