Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 1164
    Prev
    Next
    INDEX
    Explanations

    similes or comparisons

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    +0.10
    0.5%
    244
    +0.10
    0.4%
    390
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    692
    +0.10
    0.06
    512
    +0.10
    0.08
    172
    +0.09
    0.07
    Negative Logits
    erity
    -0.85
    iencies
    -0.79
     furthermore
    -0.76
     moreover
    -0.75
    icals
    -0.75
    autions
    -0.73
    yd
    -0.71
    asio
    -0.71
    hani
    -0.69
    idates
    -0.68
    POSITIVE LOGITS
     crap
    0.81
     spaghetti
    0.81
     sponge
    0.80
     Frankenstein
    0.77
     nightmare
    0.76
     pi
    0.76
     lifeless
    0.75
     encyclopedia
    0.74
     shit
    0.72
     miniature
    0.71
    Activations Density 2.291%

    No Known Activations