© 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. 49933
    Prev
    Next
    INDEX
    Explanations

    how things worked out

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

    The neuron fires on phrases that signal failure or that things “do not work out.”

    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
    legung
    -0.85
     BPM
    -0.75
     Hoof
    -0.72
    بسم
    -0.71
     tints
    -0.71
    Ứng
    -0.70
     drab
    -0.70
     DOING
    -0.69
     contribut
    -0.69
    gments
    -0.69
    POSITIVE LOGITS
     worked
    2.48
     workout
    2.27
     works
    2.09
    worked
    1.80
    Worked
    1.80
     pan
    1.73
    workout
    1.70
     Workout
    1.64
     Worked
    1.59
    Workout
    1.53
    Activations Density 0.024%

    No Known Activations