© 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. 34-GEMMASCOPE-RES-131K
    7. 870
    Prev
    Next
    INDEX
    Explanations

    film genres AND nationalities

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

    The neuron fires strongly on descriptive style and tone adjectives (and related morphemes) used in film reviews—words like “suspenseful,” “tense,” “lean,” “stripped-down,” “gothic,” etc.

    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
     humor
    -1.08
    humor
    -1.06
     sitcom
    -1.05
     comedic
    -0.96
     Humor
    -0.94
     comedians
    -0.90
     comical
    -0.89
     коми
    -0.86
     comedian
    -0.85
     hilar
    -0.84
    POSITIVE LOGITS
     drama
    1.77
     thriller
    1.51
     dramas
    1.42
     tale
    1.38
    drama
    1.23
     about
    1.21
     mystery
    1.20
     crime
    1.12
     Drama
    1.12
     adventure
    1.09
    Activations Density 0.062%

    No Known Activations