© 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 3072
    9. 8-RES_FS3072-JB
    11. 3065
    Prev
    Next
    INDEX
    Explanations

    instruction words prompting writing followed by the actual act of writing or composing 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_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    450
    +0.10
    0.5%
    377
    +0.10
    0.5%
    644
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    377
    +0.10
    0.06
    644
    +0.10
    0.05
    323
    +0.09
    0.05
    Negative Logits
    Ĭ±
    -0.93
    agara
    -0.85
    nels
    -0.70
    illon
    -0.69
    nel
    -0.68
    EGA
    -0.68
    allows
    -0.67
     Unsure
    -0.66
    xon
    -0.66
    abol
    -0.65
    POSITIVE LOGITS
     poems
    0.86
    writing
    0.86
     letters
    0.85
    smanship
    0.83
    itatively
    0.83
     memos
    0.82
     essays
    0.79
     poetry
    0.79
    acters
    0.78
     journal
    0.78
    Activations Density 0.568%

    No Known Activations