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 768
    5. 8-RES_FS768-JB
    6. 564
    Prev
    Next
    INDEX
    Explanations

    instructions and commands related to setting up and configuring software tools

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    641
    +0.14
    0.7%
    326
    +0.13
    0.6%
    167
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.14
    0.19
    641
    +0.13
    0.14
    435
    +0.12
    0.07
    Negative Logits
    borough
    -0.84
    agher
    -0.81
    lie
    -0.75
     Holt
    -0.75
    aughters
    -0.74
    abouts
    -0.74
     Hilton
    -0.74
    drug
    -0.72
     Mills
    -0.71
    onna
    -0.71
    POSITIVE LOGITS
     plugin
    1.45
     plugins
    1.40
     backend
    1.39
     SDK
    1.39
     sandbox
    1.36
     API
    1.33
     APIs
    1.33
    Version
    1.32
     debugging
    1.30
     executable
    1.29
    Activations Density 5.032%

    No Known Activations