© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129708
    Prev
    Next
    INDEX
    Explanations

    The main thing this neuron does is detect occurrences of the word “fill” (including as part of larger tokens).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    filling or completing

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ↵↵
    -3.75
    </b>
    -2.95
    <i>
    -2.64
    ↵
    -2.59
    </td>
    -2.31
    咈
    -2.23
    er
    -2.23
    <b>
    -2.22
    ay
    -2.19
    酩
    -2.19
    POSITIVE LOGITS
    我这
    2.70
    ꜆
    2.67
    </strong>
    2.67
    这件事情
    2.55
    </h3>
    2.55
    鐃
    2.44
    ၜ
    2.38
    是我的
    2.33
    
    2.30
    ༴
    2.28
    Activations Density 0.019%

    No Known Activations