© 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. 80832
    Prev
    Next
    INDEX
    Explanations

    proper nouns followed by punctuation

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

    The neuron fires most strongly on technical or domain-specific keywords—like material names (“cement”), code directives (“#include”), parameter labels (“path”), or numbered references (“lines”)—i.e. specialized jargon terms.

    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
     so
    -1.08
     epitom
    -1.07
     which
    -0.98
     grants
    -0.87
    Risultati
    -0.86
     this
    -0.85
     because
    -0.85
    timeFormat
    -0.83
     says
    -0.81
    牲
    -0.81
    POSITIVE LOGITS
     sieve
    0.96
    unless
    0.94
    ")))
    0.93
    only
    0.91
    也会
    0.90
    Articolo
    0.90
     钢
    0.90
    Occasionally
    0.88
    Unless
    0.88
    eddable
    0.88
    Activations Density 0.000%

    No Known Activations