© 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 49152
    9. 8-RES_FS49152-JB
    11. 24016
    Prev
    Next
    INDEX
    Explanations

    references to specific proper nouns, particularly names associated with crime or notable individuals

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    414
    +0.11
    0.5%
    616
    +0.11
    0.5%
    647
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.11
    0.01
    628
    +0.11
    0.01
    284
    +0.10
    0.01
    Negative Logits
    SIGN
    -0.72
    iments
    -0.71
    undai
    -0.69
    poons
    -0.69
    emade
    -0.65
     disadvant
    -0.65
    ombs
    -0.64
    ottest
    -0.64
    igree
    -0.64
    aston
    -0.63
    POSITIVE LOGITS
    brush
    1.47
     grou
    1.03
    cliffe
    0.90
    sonian
    0.87
    lings
    0.87
     sage
    0.84
    ful
    0.80
     Advice
    0.79
    vana
    0.79
    fulness
    0.79
    Activations Density 0.006%

    No Known Activations