© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 11138
    Prev
    Next
    INDEX
    Explanations

    words related to the effects of specific conditions or changes on individuals or populations

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

    after "be," "are," or "were" that describe something affected

    np_acts-logits-general · gemini-2.0-flash

    adversely affected or impacted

    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
    RRect
    -0.61
    DidLoad
    -0.56
    ImageContext
    -0.54
     Monfieur
    -0.54
     Prid
    -0.53
    Recon
    -0.53
     pleaſure
    -0.53
     myſelf
    -0.52
    __':
    -0.51
    cetype
    -0.51
    POSITIVE LOGITS
     affected
    2.27
     impacted
    1.96
     Affected
    1.89
    affected
    1.86
    Affected
    1.74
     unaffected
    1.54
     afectada
    1.45
     effected
    1.44
     afectado
    1.44
     impact
    1.36
    Activations Density 0.601%

    No Known Activations