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 1536
    5. 8-RES_FS1536-JB
    6. 1310
    Prev
    Next
    INDEX
    Explanations

    information related to events, fundraisers, and charitable activities

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    71
    +0.14
    0.6%
    723
    +0.11
    0.5%
    359
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    498
    +0.14
    0.12
    723
    +0.11
    0.09
    554
    +0.10
    -0.03
    Negative Logits
     hardness
    -0.88
     brute
    -0.81
     impe
    -0.80
    weak
    -0.80
     screws
    -0.78
     incomp
    -0.78
     overpowered
    -0.76
     weakness
    -0.75
     hairs
    -0.75
     manip
    -0.75
    POSITIVE LOGITS
     fundra
    1.41
     fundraising
    1.41
     fundraiser
    1.40
     Celebration
    1.23
     Volunteers
    1.22
     outreach
    1.17
     festival
    1.17
     charitable
    1.16
    Particip
    1.15
     nonprofit
    1.15
    Activations Density 5.704%

    No Known Activations