© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 9197
    Prev
    Next
    INDEX
    Explanations

    The neuron is especially sensitive to tokens carrying numeric or statistical data (e.g. scores, dates, measurements).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    hyphenated compound words or phrases.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    short tokens bounded by punctuation—especially hyphenated connectors (e.g., in compound words/number ranges) and standalone initials in names.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     =
    2.08
     
    2.05
    _
    1.95
     {
    1.88
    ↵
    1.86
    	
    1.84
      
    1.80
    <0x0D>
    1.77
     +
    1.75
     (
    1.71
    POSITIVE LOGITS
    ेंद्रीय
    2.09
     निजामाच्या
    2.08
     अल्को
    2.00
    प्टेंबर
    2.00
    ल्ल्या
    1.97
     आरमार
    1.97
    ॕ
    1.93
     आरमारा
    1.92
     सूत्रकृ
    1.88
     रक्षाबंधनाच्या
    1.87
    Activations Density 0.000%

    No Known Activations