Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Gemma-3-1B-IT
    3. 17-GEMMASCOPE-2-RES-16K
    4. 12616
    Prev
    Next
    INDEX
    Explanations

    This neuron detects explicit instructions and formatting or response directives in system or prompt text.

    oai_token-act-pair · gpt-5-miniTriggered by @toaster
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    gg-gs/gemma-scope-2-1b-it/resid_post
    Prompts (Dashboard)
    273,612 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    <unused1468>
    0.52
    <unused1465>
    0.51
    <unused1563>
    0.51
    <unused1438>
    0.50
    <unused1453>
    0.49
    <unused2204>
    0.49
    <unused801>
    0.49
    <unused763>
    0.49
    <unused1377>
    0.49
    <unused2197>
    0.48
    POSITIVE LOGITS
     A
    0.34
     위한
    0.32
     Do
    0.32
     어
    0.32
     Pot
    0.32
     Sa
    0.31
    이란
    0.30
     Group
    0.30
     D
    0.30
     Tool
    0.30
    Activations Density 0.021%

    No Known Activations