Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 18579
    Prev
    Next
    INDEX
    Explanations

    the word 'order' or variations like 'ordered, ordering'

    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
    512
    +0.10
    0.5%
    235
    +0.09
    0.4%
    706
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    307
    +0.10
    0.03
    133
    +0.09
    0.03
    412
    +0.09
    0.03
    Negative Logits
    tex
    -0.71
    eworld
    -0.71
    redited
    -0.69
    oat
    -0.68
    tops
    -0.68
    odd
    -0.68
    typ
    -0.66
    tek
    -0.66
    ider
    -0.66
    oos
    -0.65
    POSITIVE LOGITS
     Osw
    0.86
     guaranteeing
    0.71
     financing
    0.71
    lies
    0.70
     bolster
    0.69
     bolst
    0.67
     facilitate
    0.66
     incentiv
    0.66
     justifying
    0.66
     justify
    0.65
    Activations Density 0.017%

    No Known Activations