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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 768
    9. 8-RES_FS768-JB
    11. 761
    Prev
    Next
    INDEX
    Explanations

    sentences related to business openings or launches

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    362
    +0.21
    1.0%
    393
    +0.20
    1.0%
    608
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.21
    0.22
    393
    +0.20
    0.16
    608
    +0.13
    0.18
    Negative Logits
     besides
    -0.81
     existed
    -0.73
    rade
    -0.71
     irrespective
    -0.69
    hm
    -0.69
     relates
    -0.68
    ago
    -0.67
    each
    -0.67
     owes
    -0.67
     owing
    -0.67
    POSITIVE LOGITS
     largest
    1.39
     easiest
    1.35
     biggest
    1.35
     smallest
    1.33
     same
    1.32
     strongest
    1.32
     fastest
    1.32
     longest
    1.32
     highest
    1.31
     hottest
    1.30
    Activations Density 2.833%

    No Known Activations