© 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. 22-GEMMASCOPE-RES-131K
    7. 72971
    Prev
    Next
    INDEX
    Explanations

    saving, copying, or managing previous states

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

    The neuron strongly activates on individual identifiers or state‐describing nouns in code examples (e.g. variable names like tmp, old, saved, copy, message, time, b, country).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
    Carcinogenicity
    -0.94
    pau
    -0.92
    Açık
    -0.88
    aucune
    -0.86
    leyebilirsiniz
    -0.85
    maxAge
    -0.84
    GOV
    -0.84
     describir
    -0.82
    会う
    -0.80
    alny
    -0.80
    POSITIVE LOGITS
     some
    1.20
    韩国
    1.02
    本日
    0.96
     now
    0.94
     tadi
    0.92
     createSprite
    0.91
    ]";
    0.90
     also
    0.90
    '";
    0.88
     one
    0.87
    Activations Density 0.005%

    No Known Activations