© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 517
    Prev
    Next
    INDEX
    Explanations

    references to domestic violence and abusive behavior

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    descriptions of physical abuse, violence, and harmful behavior.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     karşılaş
    -0.35
     fewer
    -0.35
     estudian
    -0.34
    übers
    -0.34
    ustimmung
    -0.33
    accompag
    -0.32
    hopp
    -0.32
    盘
    -0.31
     automáticamente
    -0.30
    ANAL
    -0.29
    POSITIVE LOGITS
    MessageTagHelper
    0.73
     abusive
    0.66
     ब्रेकडाउन
    0.66
    олові
    0.65
    ftagPool
    0.63
     abuser
    0.62
     CreateTagHelper
    0.61
    IsMutable
    0.61
    enterOuterAlt
    0.60
    ReusableCell
    0.59
    Activations Density 0.200%

    No Known Activations