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 - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 117349
    Prev
    Next
    INDEX
    Explanations

    actions related to fighting or struggle

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     Picchu
    -0.49
    ]")]
    -0.48
    RTEX
    -0.47
    ressee
    -0.45
     arrivals
    -0.44
     dAtA
    -0.44
    convertView
    -0.43
    urfs
    -0.43
    <_>
    -0.42
     ModelRenderer
    -0.42
    POSITIVE LOGITS
     fight
    1.08
     fought
    1.02
     Fight
    1.02
     FIGHT
    1.01
    Fight
    1.01
    fight
    1.00
     fighting
    0.96
    Fighting
    0.90
    fought
    0.89
    FIGHT
    0.86
    Activations Density 0.019%

    No Known Activations