© 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. 339
    Prev
    Next
    INDEX
    Explanations

    phrases related to decision-making and personal responsibility

    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
    430
    +0.11
    0.5%
    572
    +0.10
    0.5%
    373
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.11
    0.03
    18
    +0.10
    0.03
    639
    +0.10
    0.03
    Negative Logits
    perture
    -0.64
     FINAL
    -0.64
    Wide
    -0.58
     antioxid
    -0.56
     Ratings
    -0.54
     Nile
    -0.52
     interchangeable
    -0.51
     applause
    -0.49
    WARNING
    -0.49
     Investig
    -0.49
    POSITIVE LOGITS
    nery
    1.05
    nam
    0.98
    acle
    0.95
    chard
    0.93
    Else
    0.92
    ifice
    0.90
    lando
    0.89
    ific
    0.82
    acles
    0.81
    acular
    0.80
    Activations Density 0.099%

    No Known Activations