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. 285
    Prev
    Next
    INDEX
    Explanations

    mathematical equations and notation

    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
    127
    +0.12
    0.5%
    591
    +0.10
    0.5%
    604
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    127
    +0.12
    0.10
    493
    +0.10
    0.10
    221
    +0.09
    0.07
    Negative Logits
     guiName
    -1.04
    ãĢIJ
    -0.69
    ":[{"
    -0.67
    ãĢİ
    -0.61
    wcs
    -0.57
    ifference
    -0.57
    etheless
    -0.56
     (?,
    -0.54
    htaking
    -0.53
    gencies
    -0.53
    POSITIVE LOGITS
    )
    2.20
    ?)
    2.15
    )."
    2.12
    )"
    2.08
    ),
    2.06
    !)
    2.03
    ).
    2.01
    ),"
    2.01
    *)
    1.98
    ?),
    1.98
    Activations Density 5.201%

    No Known Activations