© 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. 37334
    Prev
    Next
    INDEX
    Explanations

    references to financial expenses and costs

    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
    288
    +0.13
    0.6%
    71
    +0.12
    0.5%
    424
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    285
    +0.13
    0.01
    424
    +0.12
    0.01
    114
    +0.11
    0.01
    Negative Logits
    unin
    -0.90
    ezvous
    -0.81
    redd
    -0.77
    gewater
    -0.73
    soType
    -0.69
    ullivan
    -0.69
    ENN
    -0.68
    IPS
    -0.68
    erk
    -0.67
    erd
    -0.65
    POSITIVE LOGITS
     incurred
    1.64
     expenses
    1.11
     reimburse
    1.08
     reimb
    1.02
     reimbursement
    1.00
     expense
    0.92
     expenditure
    0.87
     ratios
    0.86
     incur
    0.86
     payable
    0.83
    Activations Density 0.011%

    No Known Activations