© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 16k
    9. 9-GEMMASCOPE-RES-16K
    11. 8716
    Prev
    Next
    INDEX
    Explanations

    references to harm and its effects

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

    "harm" or "harming"

    np_acts-logits-general · gemini-2.0-flash

    harm or hurt

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_9/width_16k/average_l0_88
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.9.hook_resid_post
    Hook Layer
    9
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    __(/*!
    -0.60
     casó
    -0.50
     recargable
    -0.46
     WebDriverWait
    -0.46
    gonic
    -0.41
     Elections
    -0.41
     Morel
    -0.40
     moods
    -0.40
     ansik
    -0.40
    itschrift
    -0.40
    POSITIVE LOGITS
     harm
    1.04
    harm
    1.00
     hurt
    1.00
    hurt
    0.95
     Harm
    0.88
     Hurt
    0.85
    Hurt
    0.85
     harmed
    0.84
     harmful
    0.81
     hurting
    0.81
    Activations Density 0.271%

    No Known Activations