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. 784
    Prev
    Next
    INDEX
    Explanations

    phrases related to philosophical and political discussions

    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
    640
    +0.35
    1.8%
    160
    +0.18
    0.9%
    107
    +0.16
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.35
    0.30
    107
    +0.18
    0.26
    160
    +0.16
    0.28
    Negative Logits
     undet
    -0.76
     neighb
    -0.74
     censored
    -0.74
     coral
    -0.74
     lifes
    -0.73
     spitting
    -0.73
     que
    -0.73
     ability
    -0.73
     bid
    -0.72
     bloc
    -0.72
    POSITIVE LOGITS
    Advertisement
    1.86
    Advertisements
    1.62
    Anyway
    1.61
    Related
    1.60
    However
    1.60
    What
    1.59
    Because
    1.58
    But
    1.58
    Unfortunately
    1.58
    Nevertheless
    1.57
    Activations Density 1.464%

    No Known Activations