© 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 49152
    9. 8-RES_FS49152-JB
    11. 37257
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "an."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    409
    +0.10
    0.5%
    755
    +0.10
    0.4%
    339
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.10
    0.01
    343
    +0.10
    0.01
    659
    +0.09
    0.01
    Negative Logits
     GEAR
    -0.77
     conclud
    -0.77
    orsi
    -0.76
    Ö¼
    -0.75
    DERR
    -0.72
    DragonMagazine
    -0.71
     NEXT
    -0.70
     destro
    -0.69
    EGIN
    -0.69
    è£ħ
    -0.67
    POSITIVE LOGITS
    ghai
    1.31
    xiety
    1.17
    ning
    1.13
    assis
    1.11
    onymous
    1.11
    ctuary
    1.06
    itary
    1.01
    ieri
    0.96
    onym
    0.95
    archy
    0.94
    Activations Density 0.012%

    No Known Activations