© 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. 22-GEMMASCOPE-RES-131K
    7. 67546
    Prev
    Next
    INDEX
    Explanations

    navigation and direction

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

    The neuron is essentially picking out code‐related tokens—identifiers, method names, punctuation, and other programming syntax elements within code blocks.

    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
    pirable
    -1.02
    werken
    -0.98
    DateTimeField
    -0.95
    ciech
    -0.95
    はる
    -0.94
    kampf
    -0.91
    erobic
    -0.91
    ETYPE
    -0.88
    Muslims
    -0.87
     глаза
    -0.87
    POSITIVE LOGITS
     direction
    1.13
     move
    1.02
     duros
    1.01
     help
    0.98
     increase
    0.96
     vuol
    0.95
    UREAU
    0.93
     appuy
    0.92
     moves
    0.91
     experiencia
    0.91
    Activations Density 0.027%

    No Known Activations