© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128691
    Prev
    Next
    INDEX
    Explanations

    The neuron activates strongly on past‐tense or past‐participle verbs (–ed forms) used as standalone headings or section‐title words.

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

    states of processes

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     vasto
    -1.44
     abbraccio
    -1.41
    などが
    -1.37
    
    -1.35
    などは
    -1.34
    
    -1.34
    玊
    -1.34
     pakan
    -1.34
    珢
    -1.30
     ekonom
    -1.30
    POSITIVE LOGITS
    -
    1.81
    In
    1.71
    Before
    1.60
    来週
    1.56
    Even
    1.52
    ことも
    1.50
     noodzake
    1.50
    At
    1.50
    L
    1.47
    久し
    1.46
    Activations Density 0.108%

    No Known Activations