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 · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 45982
    Prev
    Next
    INDEX
    Explanations

    the phrase "it turns out."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.20
    1.0%
    549
    +0.10
    0.5%
    673
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    174
    +0.20
    0.02
    673
    +0.10
    0.02
    195
    +0.10
    0.01
    Negative Logits
    lain
    -0.68
    è¦ļéĨĴ
    -0.66
     recre
    -0.59
    itation
    -0.58
     Zan
    -0.58
    riage
    -0.57
    istan
    -0.57
     Colleg
    -0.55
    ribution
    -0.55
    olin
    -0.55
    POSITIVE LOGITS
    Ī
    0.75
    ĸ
    0.70
    °
    0.64
    ij
    0.64
    quickShipAvailable
    0.62
     sour
    0.61
    nen
    0.61
     beet
    0.61
    buck
    0.60
    Ń
    0.59
    Activations Density 0.019%

    No Known Activations