© 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. 66715
    Prev
    Next
    INDEX
    Explanations

    transitioning into descriptions

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

    The neuron consistently fires on the token “to,” especially when it appears in code as part of method calls or conversions (e.g. “.toDF,” “.toJSON,” 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
    などが
    -1.62
    がなく
    -1.60
    などを
    -1.56
    ogical
    -1.52
    が
    -1.50
    titial
    -1.48
    があり
    -1.47
    鼍
    -1.43
    epar
    -1.41
     inquiet
    -1.40
    POSITIVE LOGITS
    During
    2.39
    When
    2.27
    After
    2.19
    Although
    2.08
    While
    2.03
    Shortly
    2.02
    Sometimes
    1.98
    Despite
    1.97
    Many
    1.96
    There
    1.91
    Activations Density 0.001%

    No Known Activations