© 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. 22-GEMMASCOPE-RES-131K
    7. 95360
    Prev
    Next
    INDEX
    Explanations

    error messages or failures

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

    The neuron flags punctuation and numeric tokens typical of error‐log or stack‐trace output (colons, semicolons, parentheses, question marks, error codes, 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
    mbgg
    -1.06
     amazingly
    -1.03
     strikingly
    -1.02
     remarkably
    -1.02
     faciliter
    -1.00
    unfortunately
    -0.98
     chevaux
    -0.98
    ￣)
    -0.98
     magnificent
    -0.98
     lavabo
    -0.98
    POSITIVE LOGITS
     because
    1.52
     due
    1.25
     reasons
    1.24
    理由は
    1.21
     it
    1.17
     at
    1.17
     could
    1.15
     caused
    1.14
    的原因
    1.13
     by
    1.11
    Activations Density 0.026%

    No Known Activations