© 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. 35012
    Prev
    Next
    INDEX
    Explanations

    error messages and failures

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

    The neuron homes in on error-handling code, firing strongly on tokens like “Error,” “catch,” “Exception,” and related error-message constructs.

    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
    Hence
    -1.00
     decim
    -0.93
     abſ
    -0.85
     chrétienne
    -0.85
     aquarelle
    -0.84
     Piala
    -0.83
     taupe
    -0.83
    っぱり
    -0.83
     tund
    -0.82
    Therefore
    -0.82
    POSITIVE LOGITS
     fail
    1.07
     failed
    1.07
    czenia
    1.06
     success
    1.02
     ổ
    0.98
     fails
    0.94
    ጴ
    0.92
     stone
    0.92
    Istorija
    0.92
     remained
    0.90
    Activations Density 0.001%

    No Known Activations