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. 11-GEMMASCOPE-RES-16K
    6. 1972
    Prev
    Next
    INDEX
    Explanations

    references to sequential order or hierarchical structures, particularly with the terms indicating "third" and "fourth."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Hook Layer
    11
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    WaitGroup
    -0.59
     browse
    -0.55
    oon
    -0.54
     MEC
    -0.54
     WebDriverWait
    -0.52
    CardBody
    -0.51
    \}^{
    -0.50
     ecco
    -0.50
    jestic
    -0.49
    Namara
    -0.49
    POSITIVE LOGITS
     third
    1.19
     THIRD
    1.06
     fourth
    1.05
    third
    1.01
    Third
    1.01
     Third
    0.99
     Thirdly
    0.99
     ketiga
    0.96
    THIRD
    0.94
     troisième
    0.92
    Activations Density 0.482%

    No Known Activations