© 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-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 31859
    Prev
    Next
    INDEX
    Explanations

    expressions of shame or disgrace related to social or moral issues

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

    strong negative judgment

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    Personendaten
    -0.54
     linkovi
    -0.52
     ligiloj
    -0.48
    __':
    -0.47
    +#+#
    -0.47
     otomatig
    -0.46
    ulemon
    -0.46
     seragam
    -0.46
    IntoConstraints
    -0.45
    verwijspagina
    -0.44
    POSITIVE LOGITS
    SuppressLint
    0.62
     shameful
    0.55
     appalling
    0.52
    Shame
    0.52
     disgusting
    0.52
     Shame
    0.50
    shame
    0.50
     outrageous
    0.50
     incroy
    0.49
     pathetic
    0.49
    Activations Density 0.034%

    No Known Activations