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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B SAEs
    3. Gemma-2B
    4. Residual Stream
    5. 12-RES-JB
    6. 7854
    Prev
    Next
    INDEX
    Explanations

    unusual or significant occurrences referenced in a temporal context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.12.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    12
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.34
    1.2%
    381
    +0.12
    0.4%
    690
    +0.11
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    381
    +0.34
    0.10
    1959
    +0.12
    0.14
    2044
    +0.11
    0.14
    Negative Logits
    <bos>
    -2.66
    {}
    
    -0.63
     interessa
    -0.61
    <!--
    
    -0.57
    };*/
    -0.57
    Organisateur
    -0.56
    contentLoaded
    -0.56
    ///**
    -0.56
    //*/
    -0.55
     occorre
    -0.55
    POSITIVE LOGITS
     Occasionally
    1.09
     oftentimes
    1.08
    Occasionally
    1.05
     sometimes
    1.04
     Sometimes
    1.00
     Often
    0.95
     usually
    0.94
     Usually
    0.94
     occasionally
    0.92
    sometimes
    0.92
    Activations Density 2.700%

    No Known Activations