© 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. 9837
    Prev
    Next
    INDEX
    Explanations

    phrases related to people's professions and achievements

    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
    447
    +0.23
    1.0%
    481
    +0.13
    0.6%
    353
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    761
    +0.23
    0.06
    108
    +0.13
    0.04
    70
    +0.10
    0.05
    Negative Logits
    ppings
    -0.74
    SPONSORED
    -0.71
    ulus
    -0.69
    eus
    -0.65
    rin
    -0.64
     assurances
    -0.61
    rium
    -0.60
    xit
    -0.60
    fal
    -0.59
    imaru
    -0.58
    POSITIVE LOGITS
     professionally
    0.99
     freelance
    0.98
     internship
    0.91
     gigs
    0.85
     blogging
    0.84
    chool
    0.83
     photography
    0.82
     stint
    0.79
     consultancy
    0.78
     solo
    0.74
    Activations Density 0.563%

    No Known Activations