© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 13005
    Prev
    Next
    INDEX
    Explanations

    mentions of specific measurements, statistics, and technical details related to an incident

    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
    179
    +0.11
    0.5%
    283
    +0.10
    0.5%
    707
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.11
    0.06
    640
    +0.10
    0.04
    679
    +0.09
    0.04
    Negative Logits
     nodd
    -0.69
     butcher
    -0.68
     fingert
    -0.68
     feared
    -0.67
    eled
    -0.67
     wiser
    -0.66
     wall
    -0.66
     reclaimed
    -0.65
    ding
    -0.65
     winding
    -0.64
    POSITIVE LOGITS
     Additionally
    0.96
     Therefore
    0.95
    *)
    0.95
    aspx
    0.94
     Furthermore
    0.89
     Alternatively
    0.89
     Also
    0.87
    google
    0.83
     Moreover
    0.83
     Anyway
    0.83
    Activations Density 0.077%

    No Known Activations