© 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 12288
    9. 8-RES_FS12288-JB
    11. 861
    Prev
    Next
    INDEX
    Explanations

    numbers and numerical references

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    608
    +0.10
    0.5%
    639
    +0.09
    0.4%
    490
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.10
    0.05
    312
    +0.09
    0.02
    471
    +0.09
    0.02
    Negative Logits
    oris
    -0.81
    ran
    -0.74
     echoed
    -0.72
    enery
    -0.71
    NET
    -0.71
    icipated
    -0.71
    rir
    -0.70
    ulet
    -0.70
    icious
    -0.69
    ris
    -0.69
    POSITIVE LOGITS
     rubble
    1.01
     ashes
    0.92
     Pieces
    0.85
     mere
    0.81
     earth
    0.79
     basics
    0.79
     nil
    0.79
     zero
    0.78
     conserve
    0.77
     scraps
    0.77
    Activations Density 0.173%

    No Known Activations