© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 76129
    Prev
    Next
    INDEX
    Explanations

    the superlative

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

    The neuron lights up strongest on very high‐frequency/common tokens (e.g. “the,” “s,” “best,” numbers, etc.). In other words, it detects the most common words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Ond
    -0.90
    AVIS
    -0.87
    loß
    -0.85
    という
    -0.85
    jó
    -0.82
     trase
    -0.81
    _{+}
    -0.80
     ص
    -0.79
    обу
    -0.78
     диска
    -0.77
    POSITIVE LOGITS
     bomb
    2.23
    bomb
    1.81
     bees
    1.59
     shit
    1.41
    Bomb
    1.40
     BOMB
    1.38
     pits
    1.36
     Bomb
    1.35
     bee
    1.29
    bees
    1.23
    Activations Density 0.014%

    No Known Activations