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. 22-GEMMASCOPE-RES-16K
    6. 12779
    Prev
    Next
    INDEX
    Explanations

    punctuation, particularly periods and semicolons, indicating the end of sentences or clauses

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

    preceding or related to "however"

    np_acts-logits-general · gemini-2.0-flash

    however after punctuation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     and
    -0.75
    就知道
    -0.56
     hanem
    -0.52
     וגם
    -0.51
     چقدر
    -0.50
     Fazit
    -0.50
    しくて
    -0.49
     utan
    -0.48
    干脆
    -0.48
    AnchorStyles
    -0.46
    POSITIVE LOGITS
     however
    2.61
     However
    2.39
    However
    2.33
    however
    2.28
     entanto
    2.22
     tuttavia
    1.86
     cependant
    1.77
     obstante
    1.73
     nevertheless
    1.72
    Cependant
    1.70
    Activations Density 0.393%

    No Known Activations