© 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. 628
    Prev
    Next
    INDEX
    Explanations

    comparative phrases related to consequences

    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
    373
    +0.15
    0.7%
    307
    +0.12
    0.5%
    459
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    539
    +0.15
    0.04
    307
    +0.12
    0.04
    351
    +0.11
    0.04
    Negative Logits
    umbn
    -0.76
    uto
    -0.74
    ero
    -0.73
    onomy
    -0.69
    eni
    -0.69
    illary
    -0.68
    ilic
    -0.64
    eri
    -0.64
    ista
    -0.62
    optional
    -0.60
    POSITIVE LOGITS
     anything
    1.02
     any
    0.90
     anybody
    0.82
     anywhere
    0.81
     ever
    0.78
     anyone
    0.78
     necessarily
    0.72
     usual
    0.71
     elsewhere
    0.62
     those
    0.62
    Activations Density 0.119%

    No Known Activations