© 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 768
    9. 8-RES_FS768-JB
    11. 748
    Prev
    Next
    INDEX
    Explanations

    proper names like locations, job titles, and organizations

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    7
    +0.10
    0.5%
    668
    +0.10
    0.5%
    430
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    530
    +0.10
    0.09
    430
    +0.10
    0.14
    316
    +0.10
    0.09
    Negative Logits
    Reloaded
    -0.74
    bread
    -0.66
    izations
    -0.58
    Pin
    -0.55
    icons
    -0.55
     abundance
    -0.54
     plur
    -0.54
    peat
    -0.54
     spoil
    -0.54
    izers
    -0.53
    POSITIVE LOGITS
    ;;;;;;;;;;;;
    0.81
    .;
    0.77
     Referred
    0.72
     RELEASE
    0.70
     RHP
    0.64
    ,,
    0.63
    .,
    0.62
    jri
    0.62
     EDT
    0.62
     Converted
    0.62
    Activations Density 3.819%

    No Known Activations