© 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. 9542
    Prev
    Next
    INDEX
    Explanations

    information related to research articles, academic discussions, and professional journals

    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
    393
    +0.14
    0.7%
    0
    +0.12
    0.6%
    484
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    0
    +0.14
    0.08
    484
    +0.12
    0.08
    384
    +0.09
    0.07
    Negative Logits
    ĸļ
    -0.76
    edia
    -0.72
    stown
    -0.70
    uca
    -0.67
    iless
    -0.67
    Spoiler
    -0.66
    west
    -0.65
     Arkham
    -0.65
    going
    -0.65
    rained
    -0.65
    POSITIVE LOGITS
     accessories
    1.00
     thereby
    0.91
     assorted
    0.86
     technologies
    0.81
     equivalents
    0.80
     consequently
    0.79
     caucuses
    0.78
     terminals
    0.78
     institutions
    0.75
     frameworks
    0.74
    Activations Density 0.372%

    No Known Activations