© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 43046
    Prev
    Next
    INDEX
    Explanations

    numerical values and related expressions indicating frequency or quantity

    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
    71
    +0.12
    0.6%
    756
    +0.12
    0.6%
    604
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    737
    +0.12
    0.02
    705
    +0.12
    0.02
    312
    +0.10
    0.01
    Negative Logits
    agra
    -0.79
    eka
    -0.72
    aughs
    -0.70
    antha
    -0.67
    gart
    -0.66
    rank
    -0.65
    idges
    -0.63
    edia
    -0.63
    guyen
    -0.63
    onite
    -0.62
    POSITIVE LOGITS
     except
    1.18
     imaginable
    1.12
     whatsoever
    1.12
     alike
    0.98
     thereafter
    0.94
     conceivable
    0.89
     irrespective
    0.83
    except
    0.79
     respects
    0.77
     regardless
    0.76
    Activations Density 0.045%

    No Known Activations