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. 465
    Prev
    Next
    INDEX
    Explanations

    words related to technology and manufacturing processes

    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
    63
    +0.14
    0.6%
    48
    +0.11
    0.5%
    641
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.14
    0.11
    103
    +0.11
    0.10
    48
    +0.10
    0.09
    Negative Logits
     Boone
    -0.71
     machine
    -0.68
     Machine
    -0.64
     paraly
    -0.63
     mater
    -0.62
     mate
    -0.61
     Sapphire
    -0.61
     Alchemist
    -0.61
     clipboard
    -0.59
    xus
    -0.59
    POSITIVE LOGITS
    ruits
    1.26
    ortun
    1.22
    rozen
    1.21
    letcher
    1.20
    ortunate
    1.20
    rost
    1.18
    avour
    1.16
    athom
    1.15
    ocusing
    1.14
    avorable
    1.14
    Activations Density 3.500%

    No Known Activations