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. 262
    Prev
    Next
    INDEX
    Explanations

    phrases related to music and entertainment

    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
    362
    +0.23
    1.1%
    393
    +0.20
    0.9%
    16
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.23
    0.23
    16
    +0.20
    0.14
    393
    +0.12
    0.14
    Negative Logits
    arians
    -0.74
    ional
    -0.73
    load
    -0.70
    quit
    -0.70
    pointers
    -0.69
    fn
    -0.69
    ibl
    -0.66
    nance
    -0.65
    abel
    -0.64
    handedly
    -0.64
    POSITIVE LOGITS
     midst
    1.63
     vicinity
    1.38
     meantime
    1.35
     aftermath
    1.30
     guise
    1.28
     same
    1.18
     absence
    1.16
     slightest
    1.13
     wake
    1.09
     middle
    1.08
    Activations Density 1.544%

    No Known Activations