© 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. 15347
    Prev
    Next
    INDEX
    Explanations

    concept followed by description

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

    The neuron fires on abstract “meta” nouns and noun‐phrase anchors that signal key concepts or structures in discourse—words like purpose, process, impact, ins and outs, costs, 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_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
     and
    -1.05
     beads
    -1.05
    anız
    -1.01
    Jawaban
    -0.96
    ходили
    -0.92
     Nachweis
    -0.91
     Mitarbeit
    -0.91
     והוא
    -0.90
    ariConfig
    -0.87
     botines
    -0.87
    POSITIVE LOGITS
     it
    2.78
     its
    2.64
     that
    2.33
     therein
    2.08
     thereof
    2.03
    それを
    1.77
     associated
    1.70
     how
    1.52
    それに
    1.48
     accompanying
    1.41
    Activations Density 0.103%

    No Known Activations