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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 23841
    Prev
    Next
    INDEX
    Explanations

    technical terms and concepts related to data structure and encoding

    oai_token-act-pair · gpt-4o-miniTriggered 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
    255
    +0.18
    0.8%
    737
    +0.12
    0.6%
    678
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.18
    0.04
    326
    +0.12
    0.03
    255
    +0.12
    0.04
    Negative Logits
    soDeliveryDate
    -0.87
    hur
    -0.84
    Weather
    -0.78
    tro
    -0.74
    advertising
    -0.71
    MJ
    -0.71
    taboola
    -0.71
    hran
    -0.69
    Honest
    -0.68
    fell
    -0.68
    POSITIVE LOGITS
     width
    1.03
    Offset
    1.03
    acters
    1.03
     offset
    0.96
     pointer
    0.90
     pairs
    0.90
     separated
    0.90
     integers
    0.88
     inputs
    0.88
     increments
    0.87
    Activations Density 0.123%

    No Known Activations