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. 3841
    Prev
    Next
    INDEX
    Explanations

    instances of immediate actions or reactions that lead to changes

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

    suddenness or speed

    np_acts-logits-general · gemini-2.0-flash

    immediate actions and consequences

    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
     finally
    -1.01
     lastly
    -0.98
     ultimately
    -0.96
     eventually
    -0.94
    finally
    -0.93
     gradually
    -0.89
    eventually
    -0.87
    Ultimately
    -0.83
     Ultimately
    -0.81
     finalmente
    -0.81
    POSITIVE LOGITS
    就被
    0.71
     chóng
    0.64
    WebApplication
    0.60
    EOUS
    0.54
    就把
    0.51
    fvar
    0.51
     recognizable
    0.49
    Kjelder
    0.48
    存于互联网档案馆
    0.48
     thereafter
    0.48
    Activations Density 0.296%

    No Known Activations