© 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. 808
    Prev
    Next
    INDEX
    Explanations

    mixed activities including sports terms and medical terms

    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
    603
    +0.10
    0.5%
    462
    +0.10
    0.5%
    413
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.10
    0.04
    701
    +0.10
    0.03
    373
    +0.09
    -0.02
    Negative Logits
    enthal
    -0.66
    eers
    -0.66
     footing
    -0.59
     segregation
    -0.59
     curtains
    -0.58
     Documents
    -0.58
     localization
    -0.58
    jriwal
    -0.57
     cheek
    -0.56
     adapters
    -0.56
    POSITIVE LOGITS
    adelphia
    1.16
    agall
    0.84
    ashtra
    0.83
    ebus
    0.82
    aceutical
    0.81
    phia
    0.81
    Ļ
    0.80
    auga
    0.77
    ophe
    0.76
    İ
    0.76
    Activations Density 0.222%

    No Known Activations