© 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. 11869
    Prev
    Next
    INDEX
    Explanations

    instances where something or someone has a role in a particular event or situation

    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
    745
    +0.13
    0.6%
    608
    +0.11
    0.5%
    665
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.13
    0.04
    16
    +0.11
    0.03
    589
    +0.11
    0.03
    Negative Logits
    glass
    -0.85
    uses
    -0.72
    Glass
    -0.70
    itsu
    -0.68
    ratulations
    -0.68
    ians
    -0.64
    ibl
    -0.64
    alli
    -0.64
    sticks
    -0.63
    iveness
    -0.63
    POSITIVE LOGITS
     shaping
    1.35
     determining
    1.19
    ordinate
    1.17
     facilitating
    1.17
     organising
    1.14
     constructing
    1.09
     overseeing
    1.09
     ensuring
    1.08
     influencing
    1.06
     creating
    1.05
    Activations Density 0.116%

    No Known Activations