© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 73690
    Prev
    Next
    INDEX
    Explanations

    version numbers and dates

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

    The neuron strongly activates on numeric tokens that are parts of version numbers (e.g. digits in “2.1.984”, “2.2.1”, “3.3.4” 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
    えっ
    -1.81
    づけ
    -1.74
    ほら
    -1.70
    ところに
    -1.70
    あー
    -1.66
    A
    -1.66
     $-$
    -1.58
    様な
    -1.51
    pecabe
    -1.49
    ことも
    -1.46
    POSITIVE LOGITS
    <bos>
    2.75
    もあります
    1.74
    前は
    1.66
    1
    1.63
    とはいえ
    1.59
    ちなみに
    1.58
    しかし
    1.56
    つまり
    1.56
    2
    1.55
    ？』
    1.53
    Activations Density 0.034%

    No Known Activations