© 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. 10-GEMMASCOPE-RES-131K
    7. 101660
    Prev
    Next
    INDEX
    Explanations

    word beginnings followed by 'tion'

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

    The neuron is picking up on isolated subword fragments and hyphenation/artifact tokens (i.e. pieces of words split across lines or by dashes) typical of PDF or scanned‐text extraction.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     and
    -2.59
     every
    -2.41
     after
    -2.17
     vehemently
    -2.14
     when
    -2.11
    0
    -2.11
     there
    -2.09
    Despite
    -2.08
     pertaining
    -2.08
     dozens
    -2.05
    POSITIVE LOGITS
     REALLY
    2.06
     speciale
    1.98
     revamped
    1.97
    новый
    1.96
    міні
    1.94
     !!!!
    1.94
     dando
    1.91
     pietre
    1.88
     strikingly
    1.87
    uation
    1.84
    Activations Density 0.047%

    No Known Activations