© 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. 34-GEMMASCOPE-RES-131K
    7. 73768
    Prev
    Next
    INDEX
    Explanations

    fingerprint and FBI checks

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

    The neuron consistently lights up on tokens that are either headings/titles or other “stand-out” tokens—namely capitalized words (proper nouns, section headings) and numeric/time tokens—rather than ordinary lowercase text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     🙂
    -0.76
    antie
    -0.70
    收回
    -0.70
    gtest
    -0.68
     randomness
    -0.66
    ница
    -0.65
     affinity
    -0.65
     conquête
    -0.65
     protocols
    -0.64
    baud
    -0.64
    POSITIVE LOGITS
     fingerprints
    1.70
     fingerprint
    1.60
    Finger
    1.52
     Fingerprint
    1.52
     finger
    1.52
    finger
    1.49
    Fingerprint
    1.48
     Finger
    1.40
     FBI
    1.39
    FBI
    1.36
    Activations Density 0.023%

    No Known Activations