© 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. 9210
    Prev
    Next
    INDEX
    Explanations

    retract, retraction, and related words

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

    The neuron activates on various morphological forms of the root “retract,” such as “retract,” “retraction,” “retracting,” and “retracted.”

    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
    ↵↵
    -2.84
     But
    -2.59
    {
    -2.56
    '
    -2.52
     are
    -2.47
    (
    -2.42
    ve
    -2.30
    >
    -2.27
     prioriti
    -2.27
    }$
    -2.25
    POSITIVE LOGITS
    ୌ
    2.86
    .”
    2.80
    ↵
    2.47
    ჻
    2.45
    ’
    2.38
    鰾
    2.30
    ೬
    2.25
    敍
    2.23
    ᯾
    2.20
    剛剛
    2.19
    Activations Density 0.004%

    No Known Activations