© 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. 8065
    Prev
    Next
    INDEX
    Explanations

    details and technical information

    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
    500
    +0.13
    0.6%
    40
    +0.09
    0.4%
    135
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    500
    +0.13
    0.06
    680
    +0.09
    0.04
    194
    +0.09
    0.03
    Negative Logits
     imperson
    -0.64
    odied
    -0.64
    ONES
    -0.64
     simul
    -0.64
    âĹ¼
    -0.63
     transpl
    -0.61
    pires
    -0.60
     Impossible
    -0.59
     raping
    -0.59
     salute
    -0.59
    POSITIVE LOGITS
     FAQ
    1.12
     datas
    1.05
    autions
    0.99
     wiki
    0.98
     section
    0.94
     documentation
    0.94
     appendix
    0.94
    FAQ
    0.93
     subsections
    0.91
     Definitions
    0.91
    Activations Density 0.229%

    No Known Activations