© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 128400
    Prev
    Next
    INDEX
    Explanations

    This neuron detects code identifiers and UI labels that denote creation or addition actions (e.g. tokens like “Add,” “Create,” or “New”).

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

    action verbs

    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
     فهي
    -1.05
    ischem
    -1.03
    extremely
    -0.98
    Mostly
    -0.98
    脸上的
    -0.97
     diverses
    -0.97
    roughly
    -0.96
    Appropriate
    -0.95
     zastos
    -0.94
    mainly
    -0.94
    POSITIVE LOGITS
     new
    1.70
    }$.
    1.27
    畷
    1.16
    (!__
    1.15
    ();
    1.14
     choosing
    1.07
     mengisi
    1.07
    onete
    1.06
     or
    1.05
     time
    1.05
    Activations Density 0.019%

    No Known Activations