© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 65k
    9. 16-GEMMASCOPE-RES-65K
    11. 31888
    Prev
    Next
    INDEX
    Explanations

    instances of remorse and factors influencing accountability

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot

    apology and regret

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Negative Logits
     vengeance
    -0.58
     revenge
    -0.58
     Allez
    -0.56
     emancipation
    -0.55
     empowered
    -0.55
     empowerment
    -0.55
     avenge
    -0.54
    phylaxis
    -0.54
     Pallas
    -0.54
     mola
    -0.53
    POSITIVE LOGITS
     remorse
    0.92
     apologized
    0.90
     apology
    0.88
     apologizing
    0.86
     apologize
    0.82
     apologised
    0.78
     apologe
    0.74
     apologise
    0.70
    apo
    0.70
     apolog
    0.70
    Activations Density 0.287%

    No Known Activations