© 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. 697
    Prev
    Next
    INDEX
    Explanations

    software and technology related terms

    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
    659
    +0.22
    1.1%
    67
    +0.14
    0.6%
    578
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.22
    0.11
    67
    +0.14
    0.07
    578
    +0.11
    0.05
    Negative Logits
     Chero
    -0.69
     Tinder
    -0.69
    uyomi
    -0.68
     Yamaha
    -0.67
     Eston
    -0.64
     GOODMAN
    -0.63
     Myster
    -0.60
     Gors
    -0.59
     timeout
    -0.59
     Wrestle
    -0.59
    POSITIVE LOGITS
    overed
    1.33
    URRENT
    1.27
    rossover
    1.22
    ottage
    1.20
    auldron
    1.19
    rouch
    1.18
    esar
    1.18
    ategor
    1.18
    actus
    1.17
    ursor
    1.17
    Activations Density 0.080%

    No Known Activations