© 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. 43098
    Prev
    Next
    INDEX
    Explanations

    punctuation

    np_max-act-logits · gemini-2.0-flash

    terms related to error messages or warnings in technical contexts.

    oai_token-act-pair · deepseek-v3Triggered by @finnwit

    Delimiters denote important fragments of text used in a coding or technical context, commonly surrounding error messages, commands, or specific lines of code. These fragments are often used to identify issues or describe actions in programming or technical workflows.

    eleuther_acts_top20 · deepseek-v3Triggered by @finnwit

    code

    np_max-act · deepseek-v3Triggered by @finnwit
    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
     mutual
    -0.30
     pulse
    -0.29
     term
    -0.28
     her
    -0.27
     dis
    -0.27
     pop
    -0.27
     reality
    -0.27
     exchange
    -0.26
     presence
    -0.26
     gate
    -0.26
    POSITIVE LOGITS
    The
    0.40
    An
    0.37
    You
    0.35
    you
    0.35
    This
    0.35
    In
    0.35
    (
    0.35
    L
    0.34
    First
    0.34
    Your
    0.34
    Activations Density 2.759%

    No Known Activations