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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 10215
    Prev
    Next
    INDEX
    Explanations

    references to sports events, particularly in college football

    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
    284
    +0.11
    0.5%
    571
    +0.10
    0.5%
    166
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    603
    +0.11
    0.02
    284
    +0.10
    0.02
    659
    +0.09
    0.02
    Negative Logits
    essions
    -0.70
    redit
    -0.69
    etheless
    -0.69
    cffff
    -0.68
     mitigating
    -0.66
     fragmented
    -0.65
    ulhu
    -0.64
    binary
    -0.64
     accelerated
    -0.63
    merce
    -0.63
    POSITIVE LOGITS
    mary
    1.67
    anne
    1.22
    velt
    1.14
    hill
    1.11
    mont
    1.10
    wood
    1.03
    water
    1.02
     Quartz
    1.01
     Garden
    1.01
    Rose
    0.99
    Activations Density 0.082%

    No Known Activations