© 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 12288
    9. 8-RES_FS12288-JB
    11. 1729
    Prev
    Next
    INDEX
    Explanations

    negative descriptions related to appearance or behavior

    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
    480
    +0.23
    1.1%
    71
    +0.13
    0.6%
    66
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.23
    0.04
    393
    +0.13
    0.11
    66
    +0.10
    0.05
    Negative Logits
     Murdoch
    -0.63
     Hutchinson
    -0.57
     ÂŃ
    -0.57
     Hogan
    -0.57
     marrow
    -0.54
     Sochi
    -0.53
     Cadillac
    -0.53
     mull
    -0.53
     Gut
    -0.52
     Blair
    -0.51
    POSITIVE LOGITS
    SourceFile
    0.83
    heres
    0.68
    ardless
    0.65
    xxxxxxxx
    0.65
    xxxx
    0.63
    etheless
    0.63
     yours
    0.63
     ONLY
    0.62
     toget
    0.62
    initely
    0.60
    Activations Density 2.146%

    No Known Activations