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. 90
    Prev
    Next
    INDEX
    Explanations

    verbs that suggest imagination or contemplation

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

    the word "imagine" (which appears multiple times with very high activation values).

    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
     purpoſe
    -0.62
     Efq
    -0.60
     Conſ
    -0.60
    balleur
    -0.57
     christlichen
    -0.57
    Kick
    -0.56
    Broke
    -0.56
     himſelf
    -0.56
     prière
    -0.55
     ewigen
    -0.54
    POSITIVE LOGITS
     getItemCount
    0.52
    InputTagHelper
    0.47
    imagine
    0.47
     Examine
    0.44
     magazine
    0.44
     Figura
    0.44
     Imagine
    0.43
    extAlignment
    0.42
     examine
    0.42
    杂志
    0.41
    Activations Density 0.178%

    No Known Activations