© 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. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 21470
    Prev
    Next
    INDEX
    Explanations

    data parsing and encoding

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on past‐tense or past‐participle verbs ending in “-ed,” especially ones like “parsed,” “encoded,” or “decoded.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ↵
    -3.47
    ).
    -2.53
    ↵↵
    -2.31
     «
    -2.30
    <td>
    -2.20
    ᠁
    -2.19
     were
    -2.06
     â
    -2.06
     в
    -1.95
     เรา
    -1.93
    POSITIVE LOGITS
    禵
    2.64
    栳
    2.52
    甞
    2.50
    er
    2.45
    ままで
    2.41
     identitas
    2.34
    ة
    2.25
     penyanyi
    2.19
    휀
    2.16
    This
    2.09
    Activations Density 0.013%

    No Known Activations