© 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. 9829
    Prev
    Next
    INDEX
    Explanations

    statements made by or about experts

    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
    679
    +0.13
    0.6%
    631
    +0.12
    0.6%
    88
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    88
    +0.13
    0.03
    631
    +0.12
    0.03
    568
    +0.12
    0.02
    Negative Logits
    FACE
    -0.69
    strip
    -0.69
    apo
    -0.67
     Mellon
    -0.63
     Emblem
    -0.63
    AAAA
    -0.63
     Agric
    -0.62
     Winged
    -0.62
     [|
    -0.62
     Bom
    -0.62
    POSITIVE LOGITS
    sonian
    1.02
     consulted
    1.01
     opinion
    0.88
     advise
    0.87
    udic
    0.85
    icion
    0.83
    ise
    0.82
     diagnose
    0.82
    ulously
    0.81
     interpre
    0.79
    Activations Density 0.066%

    No Known Activations