© 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 24576
    9. 8-RES_FS24576-JB
    11. 9523
    Prev
    Next
    INDEX
    Explanations

    the word "not" used in contexts of negation or opposition

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    373
    +0.12
    0.6%
    674
    +0.11
    0.5%
    745
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    226
    +0.12
    0.04
    437
    +0.11
    0.03
    651
    +0.11
    0.03
    Negative Logits
    kamp
    -0.74
    velt
    -0.72
    ixel
    -0.71
    age
    -0.70
    stru
    -0.68
    pires
    -0.68
    aspers
    -0.63
    stice
    -0.60
    FAQ
    -0.59
    orf
    -0.59
    POSITIVE LOGITS
     necessarily
    1.45
    epad
    0.92
    withstanding
    0.92
     bothering
    0.89
    icably
    0.85
     merely
    0.84
    eworthy
    0.81
     unlike
    0.81
    ifying
    0.80
    ific
    0.74
    Activations Density 0.046%

    No Known Activations