© 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. 807
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to eligibility and reimbursement in various contexts

    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
    151
    +0.11
    0.5%
    288
    +0.11
    0.5%
    496
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.11
    0.03
    288
    +0.11
    0.03
    4
    +0.10
    0.02
    Negative Logits
     Coat
    -0.68
    Else
    -0.67
    ERN
    -0.63
    aster
    -0.63
     Blow
    -0.62
    urb
    -0.60
    hed
    -0.59
     beating
    -0.58
     passionate
    -0.58
    ijk
    -0.57
    POSITIVE LOGITS
    untarily
    1.13
    igible
    0.87
    iaries
    0.81
    iciary
    0.80
     beneficiaries
    0.80
    ansas
    0.76
    Reviewer
    0.74
    rieved
    0.73
    Ohio
    0.72
    iary
    0.72
    Activations Density 0.098%

    No Known Activations