© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 84883
    Prev
    Next
    INDEX
    Explanations

    Acc words and Accumbens

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

    The neuron detects the word "Acc" at the beginning of words, particularly in contexts like "Acclaim" (video game company), "Accumulibacter" (bacteria), "Accra" (city), "Accretion" (process), and "

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    This neuron primarily detects the substring “Acc” wherever it appears (e.g., as part of names, acronyms, or word prefixes).

    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
     For
    -2.67
    }
    -2.55
    ↵↵
    -2.55
    ve
    -2.47
    </h1>
    -2.45
    </h3>
    -2.44
      
    -2.42
     hauptsächlich
    -2.38
    」,
    -2.33
    -
    -2.31
    POSITIVE LOGITS
    ון
    2.67
    ꪔ
    2.56
    یتی
    2.55
    er
    2.53
    饪
    2.45
    the
    2.41
    蠵
    2.39
     odstran
    2.34
    疍
    2.30
    up
    2.28
    Activations Density 0.002%

    No Known Activations