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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9936
    Prev
    Next
    INDEX
    Explanations

    phrases that relate to the design and purpose of products

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

    "designed"

    np_acts-logits-general · gemini-2.0-flash

    designed for purpose

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
     autorité
    -0.47
     autorità
    -0.46
    abetes
    -0.44
     antaranya
    -0.43
     resourceCulture
    -0.43
     popolari
    -0.42
    rungsseite
    -0.40
     käyt
    -0.40
     disparos
    -0.40
    adaş
    -0.39
    POSITIVE LOGITS
     designed
    0.90
    Designed
    0.86
     Designed
    0.86
     intended
    0.77
     geared
    0.68
     destined
    0.68
    designed
    0.68
    Intended
    0.67
     suited
    0.67
     Intended
    0.63
    Activations Density 0.258%

    No Known Activations