© 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 EvalsExportsinterp-engineNEWAPI 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. 4875
    Prev
    Next
    INDEX
    Explanations

    descriptions of physical objects and their characteristics

    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
    640
    +0.29
    1.4%
    6
    +0.13
    0.6%
    267
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.29
    0.11
    640
    +0.13
    0.10
    267
    +0.13
    0.08
    Negative Logits
    omics
    -0.89
     education
    -0.86
     governance
    -0.85
     counterterrorism
    -0.83
     boycot
    -0.83
     affili
    -0.82
     criminal
    -0.82
     probation
    -0.81
     absentee
    -0.80
     credential
    -0.79
    POSITIVE LOGITS
     Also
    1.31
     Alternatively
    1.26
     Unfortunately
    1.23
     Otherwise
    1.23
     However
    1.20
     Uses
    1.20
     Seems
    1.18
     Normally
    1.18
     Its
    1.17
     Specifications
    1.15
    Activations Density 0.353%

    No Known Activations