© 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. 37172
    Prev
    Next
    INDEX
    Explanations

    words that emphasize particular points or draw attention to specifics

    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
    43
    +0.11
    0.5%
    243
    +0.10
    0.5%
    725
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.11
    0.02
    243
    +0.10
    0.02
    350
    +0.10
    0.02
    Negative Logits
    alter
    -0.77
    istor
    -0.64
     Reloaded
    -0.62
    00000
    -0.62
    acker
    -0.61
    llah
    -0.60
    é¾
    -0.60
    ãĤº
    -0.60
    cli
    -0.60
    hao
    -0.60
    POSITIVE LOGITS
     egregious
    0.69
     susceptible
    0.67
     considering
    0.66
     noticeable
    0.65
    romeda
    0.63
     vulnerable
    0.63
     seasoned
    0.61
     noteworthy
    0.61
    umar
    0.61
     appreciated
    0.61
    Activations Density 0.016%

    No Known Activations