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 - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 385
    Prev
    Next
    INDEX
    Explanations

    keywords related to data attributes and API elements

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

    elements related to the attributes and metadata associated with legal documents

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

    references and legal counsel

    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-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
     vectoriales
    -0.56
    tigma
    -0.52
    cellation
    -0.51
     Hire
    -0.51
    Breeding
    -0.50
    ruce
    -0.49
    olev
    -0.49
    UMBIA
    -0.48
    copa
    -0.48
    ilibrium
    -0.48
    POSITIVE LOGITS
     زیادی
    0.76
     Administrativna
    0.72
    󠁢
    0.68
    Portale
    0.67
    too
    0.66
     enough
    0.65
    NameInMap
    0.65
     Enough
    0.64
    enough
    0.64
     thousands
    0.63
    Activations Density 0.231%

    No Known Activations