© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 57382
    Prev
    Next
    INDEX
    Explanations

    charges and convictions

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

    The neuron fires strongly on terms that name or describe legal‐trial details—like the defendant’s name, trial phase, charge degree, and other courtroom references.

    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
    Investigation
    -0.77
    pokem
    -0.77
     investigative
    -0.76
    vies
    -0.75
     outbreaks
    -0.74
     shocks
    -0.74
     reel
    -0.71
    getFile
    -0.70
    ');
    -0.69
    ม้
    -0.69
    POSITIVE LOGITS
     bribery
    1.38
     corruption
    1.30
     fraud
    1.30
     murder
    1.23
     plagiarism
    1.16
     rape
    1.11
     terrorism
    1.05
     copyright
    1.04
     bullying
    0.98
     doping
    0.96
    Activations Density 0.083%

    No Known Activations