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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 9778
    Prev
    Next
    INDEX
    Explanations

    contrasting statements or arguments

    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
    373
    +0.14
    0.6%
    163
    +0.11
    0.5%
    308
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    163
    +0.14
    0.04
    588
    +0.11
    0.04
    243
    +0.09
    0.04
    Negative Logits
    rend
    -0.97
    utra
    -0.82
    othy
    -0.76
    aim
    -0.75
    creen
    -0.74
    idelines
    -0.73
    obi
    -0.72
    emp
    -0.72
    ixel
    -0.72
    illance
    -0.72
    POSITIVE LOGITS
     though
    0.96
     curs
    0.79
     tho
    0.74
     if
    0.73
     worse
    0.72
    handed
    0.70
     seasoned
    0.69
     remotely
    0.68
     paran
    0.64
     casual
    0.62
    Activations Density 0.047%

    No Known Activations