© 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. 129113
    Prev
    Next
    INDEX
    Explanations

    Explanation of neuron 4 behavior: the main thing this neuron does is find strongly positive or promotional words that highlight achievements, success, or progress.

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

    future for elite few

    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
     speculations
    -1.09
     会場
    -1.02
    ค่ะ
    -1.02
     scanty
    -1.00
     besonders
    -0.99
    錚
    -0.97
     ressource
    -0.94
     Pitcher
    -0.93
     omnibus
    -0.91
     außerdem
    -0.91
    POSITIVE LOGITS
    Plenty
    1.16
    Приятного
    1.11
    ]=(
    1.07
     mahd
    1.07
    Thankfully
    1.06
     plenty
    1.05
     =>
    1.05
    恥ず
    1.04
     personnel
    1.02
     ESPN
    1.00
    Activations Density 0.099%

    No Known Activations