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. 5193
    Prev
    Next
    INDEX
    Explanations

    arguments presented in support of a claim or position

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

    certainty or strong affirmation

    np_acts-logits-general · gemini-2.0-flash

    absolutely completely

    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
     occasionally
    -1.05
     somewhat
    -1.02
     sometimes
    -1.00
     perhaps
    -0.97
     Occasionally
    -0.96
    Occasionally
    -0.93
     occasional
    -0.92
     Sometimes
    -0.92
     maybe
    -0.92
    somewhat
    -0.91
    POSITIVE LOGITS
     absolutely
    1.50
     utterly
    1.38
    absolutely
    1.36
     completely
    1.36
     totally
    1.28
     absolutamente
    1.27
    atissimo
    1.26
     Absolutely
    1.23
    Absolutely
    1.23
     unbelievably
    1.22
    Activations Density 2.140%

    No Known Activations