© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 66994
    Prev
    Next
    INDEX
    Explanations

    code declarations

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

    The neuron activates on any token that contains a literal “.” character (e.g. file‐paths like “./”, dotted names like “config.” or “models.”, import modules, etc.).

    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
     célè
    -1.16
     phénom
    -1.13
     rafra
    -1.10
     Souha
    -1.10
     réinitial
    -1.09
     quelques
    -1.07
     ralenti
    -1.02
     både
    -1.02
     Propriété
    -0.99
    
    -0.98
    POSITIVE LOGITS
     this
    0.99
     Cycle
    0.98
    GenerationType
    0.96
     _
    0.96
    $
    0.94
    j
    0.92
    mini
    0.92
     marine
    0.90
    ju
    0.89
     without
    0.89
    Activations Density 0.052%

    No Known Activations