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. 0-RES-JB
    6. 47
    Prev
    Next
    INDEX
    Explanations

    terms related to architectural elements

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1472
    +0.10
    0.3%
    1306
    +0.08
    0.2%
    1331
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1896
    +0.10
    0.02
    1331
    +0.08
    0.03
    1343
    +0.08
    0.03
    Negative Logits
    <bos>
    -0.68
     viena
    -0.62
     tiks
    -0.59
     veik
    -0.59
     papild
    -0.55
    ///**
    -0.52
    ConverterFactory
    -0.52
     didel
    -0.51
     įsi
    -0.50
     weten
    -0.49
    POSITIVE LOGITS
    bal
    2.18
     bal
    2.05
    Bal
    1.80
     Ballot
    1.71
     Bal
    1.69
    BAL
    1.67
     BAL
    1.65
     Ballo
    1.56
     Balo
    1.52
     ballot
    1.50
    Activations Density 0.239%

    No Known Activations