© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 16k
    9. 18-GEMMASCOPE-RES-16K
    11. 6851
    Prev
    Next
    INDEX
    Explanations

    the beginning of a new section or topic in a document

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

    Interjections and conversational fillers

    np_acts-logits-general · gemini-2.0-flash

    text around punctuation

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron primarily detects the beginning-of-sequence (document start) token.

    oai_token-act-pair · gpt-5-miniTriggered by @igorkuztech
    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
    ="@+
    -0.51
    angkan
    -0.50
    ChildScrollView
    -0.49
    Nevertheless
    -0.45
     setIs
    -0.44
    それでは
    -0.44
     Therefore
    -0.44
    さすが
    -0.43
     therefore
    -0.42
    WRENCE
    -0.42
    POSITIVE LOGITS
    Oh
    0.87
     Oh
    0.85
    oh
    0.84
     CreateTagHelper
    0.82
     oh
    0.79
     BoxDecoration
    0.74
     uh
    0.74
    Personensuche
    0.73
    :✨
    0.72
     Uh
    0.72
    Activations Density 0.131%

    No Known Activations