© 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. 10-GEMMASCOPE-RES-131K
    7. 90327
    Prev
    Next
    INDEX
    Explanations

    gunman or driver manager

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

    This neuron activates on words containing the substring “man” (including its plural “men”) or the similar “ton” ending, such as “gunman,” “gunmen,” and “Middleton.”

    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
    </h3>
    -2.97
    a
    -2.61
    {
    -2.53
    A
    -2.41
    }
    -2.39
    ↵
    -2.30
    over
    -2.25
    ⻬
    -2.23
    」
    -2.22
    The
    -2.19
    POSITIVE LOGITS
    篳
    2.52
    ᝢ
    2.41
    。】
    2.41
    晛
    2.36
    膂
    2.28
    ᭝
    2.27
    翾
    2.27
    礽
    2.19
    媼
    2.17
     thinks
    2.16
    Activations Density 0.001%

    No Known Activations