© 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. 23992
    Prev
    Next
    INDEX
    Explanations

    instances of the word "and" in conjunction with various other terms

    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
    430
    +0.10
    0.5%
    247
    +0.10
    0.4%
    407
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    363
    +0.10
    0.03
    546
    +0.10
    0.03
    617
    +0.10
    0.03
    Negative Logits
    ocious
    -0.74
    bole
    -0.71
     sag
    -0.69
    agnetic
    -0.65
    utenberg
    -0.64
     shif
    -0.61
     cig
    -0.59
    ek
    -0.59
     manif
    -0.58
    wake
    -0.58
    POSITIVE LOGITS
     USSR
    0.96
    SEA
    0.95
     Europe
    0.93
     NATO
    0.93
     Allies
    0.92
    NAT
    0.91
    EU
    0.91
    Mexico
    0.87
     European
    0.84
     Soviets
    0.83
    Activations Density 0.135%

    No Known Activations