© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 6932
    Prev
    Next
    INDEX
    Explanations

    Feature comparison table

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

    This neuron detects news‐style dateline location tokens (city and region names often appearing in all caps or as proper nouns at the start of an article).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ニュース
    0.58
    Aire
    0.55
     फिज
    0.53
     पंजाबी
    0.50
    ←
    0.50
     thefts
    0.49
    ългар
    0.48
    장에서
    0.48
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    0.48
     इजी
    0.48
    POSITIVE LOGITS
     {_
    0.59
     dijo
    0.59
     loos
    0.59
     जिनके
    0.57
     ပြ
    0.57
     Primer
    0.57
     SERVER
    0.57
     (_
    0.56
     Chrom
    0.56
     bandwidth
    0.56
    Activations Density 0.091%

    No Known Activations