© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 2341
    Prev
    Next
    INDEX
    Explanations

    expressions related to struggle, hardship, or resilience

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

    highly emotional or hyperbolic language, often involving extreme actions or dramatic statements.

    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
     postup
    -0.40
     diperhatikan
    -0.38
     overall
    -0.38
     relatively
    -0.38
     mengal
    -0.38
    wnież
    -0.38
     wspól
    -0.37
     Overall
    -0.37
     önemli
    -0.37
    Overall
    -0.36
    POSITIVE LOGITS
    OGND
    0.79
     Infórmanos
    0.78
     ویکی‌پدی
    0.73
     мәкал
    0.65
     Мексичка
    0.65
     Савезне
    0.63
    sidemargin
    0.62
     shit
    0.58
    XmlIgnore
    0.58
     crap
    0.56
    Activations Density 0.185%

    No Known Activations