© 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. 22-GEMMASCOPE-RES-131K
    7. 38277
    Prev
    Next
    INDEX
    Explanations

    pre- and post- transitions

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

    The neuron fires on the core nouns in hyphenated or prefixed stage/time‐phase terms (e.g. the “training” in pre‐training, the “assessment” in post‐assessment, the “excision” in pre‐ and post‐excision, 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_22/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
     one
    -1.20
    blos
    -1.00
    えっ
    -0.98
     šo
    -0.94
    /#{
    -0.91
     venezolano
    -0.90
     یکی
    -0.90
    cartel
    -0.90
     przede
    -0.89
     musim
    -0.88
    POSITIVE LOGITS
    чники
    1.12
    <bos>
    1.10
     activities
    1.09
     strategy
    1.06
     assorted
    1.03
     cherchez
    1.02
    owner
    0.99
     Lakukan
    0.99
     sizzling
    0.98
     scenario
    0.97
    Activations Density 0.111%

    No Known Activations