© 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. 22132
    Prev
    Next
    INDEX
    Explanations

    references to explosive devices or explosions

    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
    288
    +0.15
    0.7%
    326
    +0.10
    0.5%
    34
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.15
    0.02
    34
    +0.10
    0.01
    326
    +0.10
    0.01
    Negative Logits
    heit
    -0.88
    esan
    -0.84
    cedented
    -0.77
    egal
    -0.77
    naire
    -0.77
    agall
    -0.77
    SEA
    -0.76
    pai
    -0.74
    alian
    -0.74
    atche
    -0.73
    POSITIVE LOGITS
     residue
    0.86
     decomp
    0.81
     eru
    0.79
    iating
    0.76
     deton
    0.74
     belts
    0.74
    ooter
    0.73
     explosive
    0.73
     disposal
    0.70
     incendiary
    0.68
    Activations Density 0.022%

    No Known Activations