© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 68681
    Prev
    Next
    INDEX
    Explanations

    familiarity and expectedness

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

    The neuron strongly activates on mentions of “familiar” (and related forms like “familiarity”), effectively spotting references to familiarity.

    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
    蹰
    -3.39
    萇
    -2.86
    瘊
    -2.64
     marvelous
    -2.59
    鍺
    -2.58
     různých
    -2.55
     těch
    -2.52
    
    -2.52
     tumultuous
    -2.48
    歘
    -2.48
    POSITIVE LOGITS
     Surprisingly
    2.95
    Surprisingly
    2.69
    Even
    2.69
     Basically
    2.67
     Following
    2.66
     Almost
    2.66
     That
    2.61
     Taking
    2.59
     Before
    2.58
     Making
    2.56
    Activations Density 0.006%

    No Known Activations