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. 18-GEMMASCOPE-RES-16K
    6. 3638
    Prev
    Next
    INDEX
    Explanations

    references to heraldry or family crests

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

    Family crests and heraldry

    np_acts-logits-general · gemini-2.0-flash

    family arms and heraldry

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    stylers
    -0.63
     distanciation
    -0.58
     myſelf
    -0.57
    ilight
    -0.55
     **/
    
    -0.54
     дописавши
    -0.54
    iastes
    -0.53
     domestiques
    -0.53
     filtrate
    -0.53
     Référence
    -0.52
    POSITIVE LOGITS
     herald
    0.68
     Herald
    0.66
     heral
    0.64
     Arms
    0.62
     arms
    0.61
    Herald
    0.60
     Crest
    0.58
    Arms
    0.57
    setVerticalGroup
    0.54
    herald
    0.53
    Activations Density 0.257%

    No Known Activations