© 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. 10-GEMMASCOPE-RES-131K
    7. 73407
    Prev
    Next
    INDEX
    Explanations

    combining technical components and data

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

    The neuron is keyed to the verb “combine,” strongly activating whenever that exact word appears.

    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
    l
    -3.47
    -3.13
     '
    -3.05
     
    -3.05
    r
    -2.84
    mathrm
    -2.75
     they
    -2.69
     same
    -2.64
     The
    -2.63
    si
    -2.61
    POSITIVE LOGITS
     faculté
    2.91
    dzić
    2.72
    最后由
    2.64
    以上です
    2.53
    というわけで
    2.52
     eivät
    2.45
     祝福
    2.42
    ୌ
    2.39
    先は
    2.39
     點
    2.38
    Activations Density 0.021%

    No Known Activations