Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 275
    Prev
    Next
    INDEX
    Explanations

    descriptions or references to world records or superlatives

    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
    558
    +0.10
    0.5%
    288
    +0.10
    0.4%
    376
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.16
    67
    +0.10
    0.12
    752
    +0.09
    0.11
    Negative Logits
    ippi
    -1.04
    veyard
    -0.79
    etsk
    -0.74
    sbm
    -0.73
    iple
    -0.73
    ibility
    -0.71
    edin
    -0.71
    imon
    -0.70
    ients
    -0.67
    »Ĵ
    -0.66
    POSITIVE LOGITS
    wide
    1.09
     warming
    0.90
     premiere
    0.85
    wake
    0.85
    warming
    0.84
     Wide
    0.82
     wide
    0.81
    building
    0.80
    liness
    0.80
     renowned
    0.80
    Activations Density 3.383%

    No Known Activations