Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 115
    Prev
    Next
    INDEX
    Explanations

    conclusive phrases indicating the end or summary of a discussion

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

    adverbs signaling conclusion or final points, like "finally" and "lastly".

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     liền
    -0.37
    ientôt
    -0.36
     concurrently
    -0.35
     Comprometido
    -0.35
    GOTREF
    -0.34
    是有
    -0.34
     immediately
    -0.33
    insatz
    -0.33
     Lineage
    -0.33
     склад
    -0.32
    POSITIVE LOGITS
     lastly
    0.85
     Lastly
    0.84
     Finally
    0.83
     infine
    0.81
    Finally
    0.81
    Lastly
    0.78
     finally
    0.78
    finally
    0.74
    最後に
    0.72
     FINALLY
    0.70
    Activations Density 0.161%

    No Known Activations