© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 71085
    Prev
    Next
    INDEX
    Explanations

    song was, Marsh continued, case that, problem is

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

    The neuron fires on the main topic word or “subject” noun of a new sentence or paragraph—i.e. the key content word that introduces what the sentence is about.

    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
    Recently
    -1.08
     Recently
    -1.05
     when
    -1.05
     kebersihan
    -1.03
     for
    -1.02
     or
    -0.94
     recently
    -0.93
     which
    -0.93
     recientemente
    -0.93
     vervol
    -0.92
    POSITIVE LOGITS
    の特徴
    1.08
     razones
    0.96
    chließend
    0.93
    orig
    0.90
     horrid
    0.90
    牟
    0.89
    јани
    0.89
     picturesque
    0.88
     philosoph
    0.88
     also
    0.87
    Activations Density 0.069%

    No Known Activations