© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 8853
    Prev
    Next
    INDEX
    Explanations

    octa-, in-, the-, ;-, followed by word fragments

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

    This neuron responds to individual alphanumeric characters (letters or digits) when they appear as standalone tokens—most notably inside regex character‐class definitions.

    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
    ]
    -2.03
    }
    -1.95
    s
    -1.84
     «
    -1.73
    a
    -1.69
     for
    -1.66
    "
    -1.57
    ],
    -1.52
     if
    -1.51
    o
    -1.48
    POSITIVE LOGITS
    .“
    1.98
    ጴ
    1.66
     mattino
    1.65
     Giugno
    1.60
    妡
    1.60
    enares
    1.57
    rédients
    1.57
    ቼ
    1.55
    籀
    1.53
    
    1.52
    Activations Density 0.001%

    No Known Activations