© 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. 100396
    Prev
    Next
    INDEX
    Explanations

    human leukocyte antigen (HLA)

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

    location names, particularly cities like "Newark" and acronyms like "HLA" and references to nationalities like "American Samoa".

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

    The neuron strongly responds to occurrences of the place name “Newark.”

    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
    ࿆
    -3.33
    </h3>
    -2.69
    珦
    -2.55
    ﮏ
    -2.52
    CHREIB
    -2.50
    żeli
    -2.42
    緜
    -2.39
     nächsten
    -2.38
     bezoekers
    -2.36
    𖣘
    -2.33
    POSITIVE LOGITS
    a
    3.88
    o
    2.86
    e
    2.83
    d
    2.72
    m
    2.41
    鶫
    2.19
    ﺐ
    2.11
    鯢
    2.08
    for
    2.06
    A
    2.06
    Activations Density 0.021%

    No Known Activations