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. 12701
    Prev
    Next
    INDEX
    Explanations

    definitive answers to questions, particularly focusing on affirmatives and negatives

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

    Following "is," "answered," or "said" and affirming "yes."

    np_acts-logits-general · gemini-2.0-flash

    detecting yes or no answers

    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
     BoxDecoration
    -0.80
    berdayakan
    -0.67
    WireFormatLite
    -0.67
    WebElementEntity
    -0.67
    SupportActionBar
    -0.64
     ModelExpression
    -0.64
    getDoctrine
    -0.63
     Majefty
    -0.61
    neſs
    -0.60
    Дереккөздер
    -0.60
    POSITIVE LOGITS
     yes
    1.35
     YES
    1.24
    yes
    1.16
     Yes
    1.12
    Yes
    1.11
    YES
    1.11
     affirmative
    0.94
     resounding
    0.82
     NO
    0.80
     Nope
    0.78
    Activations Density 0.227%

    No Known Activations