© 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. 22346
    Prev
    Next
    INDEX
    Explanations

    references to border enforcement and security agencies

    oai_token-act-pair · gpt-4o-miniTriggered 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
    447
    +0.14
    0.6%
    617
    +0.12
    0.6%
    5
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.14
    0.04
    617
    +0.12
    0.04
    708
    +0.11
    0.03
    Negative Logits
    jad
    -0.76
    ophy
    -0.70
    wcs
    -0.68
    oted
    -0.68
    esta
    -0.67
    etus
    -0.67
    isting
    -0.66
    eless
    -0.66
    plate
    -0.65
    heid
    -0.65
    POSITIVE LOGITS
     Officers
    1.07
     Investigations
    1.01
     troopers
    0.98
     checkpoints
    0.95
     officers
    0.93
     Agents
    0.93
     Service
    0.92
     trooper
    0.90
     personnel
    0.89
     Officer
    0.89
    Activations Density 0.070%

    No Known Activations