Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 5235
    Prev
    Next
    INDEX
    Explanations

    conversational phrases and expressions indicating surprise or acknowledgment

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

    "That" followed by "is," "was," or "'s"

    np_acts-logits-general · gemini-2.0-flash

    that's a

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    勿論
    -0.60
    ]-->
    -0.57
     természetes
    -0.53
     permitan
    -0.53
     oczywiście
    -0.52
     traditionnels
    -0.52
     především
    -0.52
     hjär
    -0.51
     uniquement
    -0.51
     cél
    -0.51
    POSITIVE LOGITS
     huh
    0.82
     impressive
    0.79
     Impressive
    0.78
     lotta
    0.77
     mighty
    0.76
    
    0.73
     tough
    0.73
     mouthful
    0.72
    Impressive
    0.71
     sacré
    0.71
    Activations Density 0.222%

    No Known Activations