© 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. 10300
    Prev
    Next
    INDEX
    Explanations

    terms related to structural plans and safety measures in high-rise buildings

    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
    288
    +0.17
    0.8%
    314
    +0.11
    0.5%
    323
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    12
    +0.17
    0.03
    476
    +0.11
    0.02
    103
    +0.10
    0.03
    Negative Logits
    RANT
    -0.74
    ously
    -0.72
     Sunder
    -0.66
     maker
    -0.64
     Bard
    -0.64
     Maker
    -0.63
     bent
    -0.62
     Slug
    -0.62
    eers
    -0.61
    eer
    -0.61
    POSITIVE LOGITS
    cend
    1.26
    mitt
    1.22
    gression
    1.16
    istors
    1.14
    lator
    1.13
    mission
    1.13
    continental
    1.12
    lucent
    1.10
    latable
    1.10
    parency
    1.08
    Activations Density 0.034%

    No Known Activations