© 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. 60438
    Prev
    Next
    INDEX
    Explanations

    Vector direction axis

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

    The neuron is detecting programming‐style tokens—especially code symbols, numeric literals, and class/constructor names (e.g. hyphens, digits, and identifiers like Vector3).

    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
    𝗹
    -0.79
    getCount
    -0.75
    ₪
    -0.73
     posicion
    -0.73
     äls
    -0.73
    alih
    -0.73
    จุด
    -0.73
     igen
    -0.73
     Deal
    -0.71
    пры
    -0.71
    POSITIVE LOGITS
     cross
    1.48
     direction
    1.36
     axis
    1.32
    cross
    1.31
     UP
    1.24
     directions
    1.24
     CROSS
    1.22
    Cross
    1.16
    方向
    1.14
    CROSS
    1.13
    Activations Density 0.012%

    No Known Activations