© 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. 30-TRANSCODER-HP
    7. 96682
    Prev
    Next
    INDEX
    Explanations

    execution

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     zero
    -0.30
    鼶
    -0.29
    (BigInteger
    -0.28
    à¹Īาว
    -0.27
    åİĭ
    -0.26
     sleep
    -0.26
    å£ĵ
    -0.25
    æķ¬è¯·
    -0.25
    oud
    -0.25
    zero
    -0.25
    POSITIVE LOGITS
     for
    0.26
    for
    0.25
    agog
    0.25
    åįĩ级
    0.25
    çĸĻ
    0.24
    è¢ĭ
    0.24
     Semantic
    0.24
     allocator
    0.24
    主å¸Ń
    0.24
    çĵ¶
    0.23
    Activations Density 0.102%

    No Known Activations