© 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. 24271
    Prev
    Next
    INDEX
    Explanations

    references to police officers and their actions

    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
    289
    +0.13
    0.6%
    567
    +0.12
    0.5%
    762
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    509
    +0.13
    0.03
    75
    +0.12
    0.03
    533
    +0.10
    0.03
    Negative Logits
     Remastered
    -0.80
    DAY
    -0.79
    âĶģ
    -0.76
    iens
    -0.74
    hered
    -0.72
    ãĤ¤ãĥĪ
    -0.70
    Shock
    -0.69
    dos
    -0.69
    ãĤ¦
    -0.69
    ãĥ¼ãĥĨ
    -0.68
    POSITIVE LOGITS
     stationed
    1.09
     officers
    1.03
     patrolling
    0.99
     officer
    0.97
    utenant
    0.95
    geant
    0.88
     tasked
    0.86
     corps
    0.85
    involved
    0.82
     Officers
    0.82
    Activations Density 0.062%

    No Known Activations