© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 37309
    Prev
    Next
    INDEX
    Explanations

    articles, particularly "a" and "an"

    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
    36
    +0.21
    1.1%
    393
    +0.17
    0.8%
    362
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.21
    0.07
    328
    +0.17
    0.04
    362
    +0.15
    0.04
    Negative Logits
    Attempts
    -0.84
    words
    -0.83
    evidence
    -0.79
    edit
    -0.77
     Accounts
    -0.76
    agents
    -0.73
    Types
    -0.72
    books
    -0.71
    Access
    -0.69
    external
    -0.68
    POSITIVE LOGITS
     bunch
    0.93
     handful
    0.85
     lot
    0.82
     stationary
    0.80
     dozen
    0.78
     peac
    0.77
     piece
    0.76
     skysc
    0.75
     whole
    0.74
     semi
    0.73
    Activations Density 0.097%

    No Known Activations