© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 8-GEMMASCOPE-RES-16K
    11. 6947
    Prev
    Next
    INDEX
    Explanations

    instances of the word "weak" and its related phrases, indicating an emphasis on weakness or inadequacy

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 8-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_8/width_16k/average_l0_99
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    encodeWith
    -0.53
    Kjelder
    -0.49
    extAlignment
    -0.48
     gainera
    -0.48
     GenerationType
    -0.47
     насељу
    -0.46
    XmlAccessorType
    -0.46
    irtschaft
    -0.46
    NamedQueries
    -0.46
    chartInstance
    -0.45
    POSITIVE LOGITS
     weak
    2.39
    weak
    2.16
     Weak
    2.13
    Weak
    2.09
     weaker
    1.73
     weakly
    1.73
     débil
    1.70
     weakness
    1.59
     debole
    1.55
     weakest
    1.55
    Activations Density 0.331%

    No Known Activations