© 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. 34-GEMMASCOPE-RES-131K
    4. 116948
    Prev
    Next
    INDEX
    Explanations

    hang and related prefixes

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

    The neuron detects occurrences of the root “hang” (i.e. words beginning with “hang,” such as hangouts, hanging, hangover, 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_34/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
     deleteById
    -0.76
     ehe
    -0.76
     getAge
    -0.75
    コント
    -0.73
     oficialmente
    -0.73
     accordingly
    -0.71
     commonly
    -0.71
    Tach
    -0.69
    wong
    -0.69
    feitura
    -0.68
    POSITIVE LOGITS
     reutiliz
    0.84
     Hang
    0.84
     Houghton
    0.81
    developing
    0.81
    Education
    0.75
    })+
    0.75
     Goddard
    0.74
    ecka
    0.73
    helper
    0.73
    tags
    0.73
    Activations Density 0.012%

    No Known Activations