© 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 24576
    9. 8-RES_FS24576-JB
    11. 18863
    Prev
    Next
    INDEX
    Explanations

    technical terms and abbreviations related to technology products and brands

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    314
    +0.16
    0.7%
    83
    +0.13
    0.6%
    627
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    537
    +0.16
    0.04
    124
    +0.13
    0.04
    314
    +0.10
    0.01
    Negative Logits
     bounce
    -0.78
     hiber
    -0.74
     Blizz
    -0.71
     Arabian
    -0.69
     darts
    -0.68
     square
    -0.67
     wheel
    -0.66
     brist
    -0.66
     scrim
    -0.65
     drink
    -0.64
    POSITIVE LOGITS
    III
    1.11
    FM
    1.09
    style
    1.04
    derived
    1.03
    NET
    1.02
    II
    1.02
    GO
    1.01
    induced
    1.01
    IV
    1.00
    RN
    0.99
    Activations Density 0.054%

    No Known Activations