© 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. 72533
    Prev
    Next
    INDEX
    Explanations

    uses 'like' for comparison

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

    The neuron fires strongly on the word “like” (especially in “just like” or “like X, …”) when it introduces a simile or comparison.

    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
     őket
    -1.01
     AssemblyTitle
    -0.97
    ulink
    -0.94
    xia
    -0.94
    बि
    -0.94
    𝕭
    -0.91
    кса
    -0.90
    РС
    -0.90
     visto
    -0.90
    ]+\
    -0.90
    POSITIVE LOGITS
     other
    1.50
     also
    1.49
     these
    1.32
    Also
    1.22
     want
    1.21
     this
    1.20
     each
    1.17
     when
    1.14
    Other
    1.08
     arī
    1.07
    Activations Density 0.014%

    No Known Activations