© 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. 1429
    Prev
    Next
    INDEX
    Explanations

    phrases related to naming conventions and origins

    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
    373
    +0.16
    0.7%
    362
    +0.11
    0.5%
    296
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.16
    0.04
    647
    +0.11
    0.03
    669
    +0.10
    0.03
    Negative Logits
    asive
    -0.84
    emn
    -0.78
    reasonable
    -0.77
    wd
    -0.76
    pour
    -0.76
    ior
    -0.76
    gencies
    -0.74
    ggle
    -0.73
    enaries
    -0.73
    isans
    -0.73
    POSITIVE LOGITS
     initials
    1.08
     slang
    1.00
     resemblance
    0.93
    ãĤ´
    0.87
     Latin
    0.86
     famous
    0.85
     plural
    0.84
     Greek
    0.83
     Anglo
    0.83
     Norse
    0.82
    Activations Density 0.228%

    No Known Activations