© 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. 729
    Prev
    Next
    INDEX
    Explanations

    instances where the concept of "purpose" is mentioned

    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
    509
    +0.12
    0.6%
    332
    +0.11
    0.5%
    86
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    509
    +0.12
    0.03
    737
    +0.11
    0.03
    688
    +0.10
    0.03
    Negative Logits
     weeds
    -0.69
     juven
    -0.64
     slopes
    -0.63
    thumbnails
    -0.63
     varieties
    -0.63
     Leopard
    -0.62
     Toad
    -0.62
     ranks
    -0.59
     Isles
    -0.59
     fists
    -0.58
    POSITIVE LOGITS
    ful
    1.31
    fully
    1.06
    fulness
    1.05
    lessly
    1.00
    lessness
    0.95
    FUL
    0.91
    tenance
    0.90
    teness
    0.89
    ality
    0.86
    bringer
    0.84
    Activations Density 0.077%

    No Known Activations