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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 3712
    Prev
    Next
    INDEX
    Explanations

    expressions of capability and success in achieving tasks or goals

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

    "to" followed by a verb

    np_acts-logits-general · gemini-2.0-flash

    managed to persuade or find

    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
    ClientSize
    -0.57
    styleType
    -0.54
    essayer
    -0.54
    出版年
    -0.53
     <<<<<<<<<<<<<<
    -0.51
     disfru
    -0.50
    Autoritní
    -0.50
     gärna
    -0.50
     Chrift
    -0.50
    potential
    -0.49
    POSITIVE LOGITS
     convince
    1.18
     persuade
    1.12
     persuading
    1.07
     convincing
    1.03
     persuaded
    1.02
     find
    0.92
     convinces
    0.90
     convencer
    0.90
     convaincre
    0.90
     convinced
    0.90
    Activations Density 0.480%

    No Known Activations