© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1899
    Prev
    Next
    INDEX
    Explanations

    names of persons being quoted or referred to in text

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    316
    +0.12
    0.5%
    121
    +0.11
    0.5%
    548
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    316
    +0.12
    0.06
    502
    +0.11
    0.05
    659
    +0.10
    0.06
    Negative Logits
    ardless
    -0.76
     disproportion
    -0.73
    paralle
    -0.69
    etheless
    -0.68
    acters
    -0.64
    noon
    -0.63
     impair
    -0.61
     disrupting
    -0.61
    ĵĺ
    -0.59
    blem
    -0.58
    POSITIVE LOGITS
     said
    1.40
     says
    1.23
     wrote
    1.22
     told
    1.21
     exclaimed
    1.19
    Said
    1.14
     replied
    1.10
    said
    1.08
     explained
    1.07
     Said
    1.06
    Activations Density 0.288%

    No Known Activations