© 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 768
    9. 8-RES_FS768-JB
    11. 646
    Prev
    Next
    INDEX
    Explanations

    instances of the word "get" followed by a number

    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
    314
    +0.19
    0.9%
    183
    +0.11
    0.5%
    372
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    183
    +0.19
    0.10
    372
    +0.11
    0.09
    62
    +0.10
    0.09
    Negative Logits
     Archdemon
    -0.64
    velt
    -0.60
     endeavour
    -0.59
     Britann
    -0.57
     experiment
    -0.57
    haus
    -0.57
     annex
    -0.57
    iege
    -0.57
     endeavor
    -0.57
     cardinal
    -0.54
    POSITIVE LOGITS
     rid
    1.43
    tin
    1.00
     acquainted
    0.98
    cloneembedreportprint
    0.97
    TING
    0.92
    DragonMagazine
    0.88
     distracted
    0.88
     bored
    0.85
     sucked
    0.82
    bent
    0.82
    Activations Density 2.636%

    No Known Activations