© 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. 793
    Prev
    Next
    INDEX
    Explanations

    mentions of a specific word: "fellow"

    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
    167
    +0.11
    0.5%
    109
    +0.09
    0.4%
    85
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    555
    +0.11
    0.02
    242
    +0.09
    0.02
    85
    +0.09
    0.02
    Negative Logits
    anwhile
    -0.68
     Zup
    -0.63
    uilt
    -0.62
     Sue
    -0.61
    wark
    -0.60
     Retrieved
    -0.60
     livest
    -0.59
     annexed
    -0.59
    onic
    -0.59
     definitively
    -0.58
    POSITIVE LOGITS
    worldly
    1.09
     travelers
    1.06
     traveler
    1.04
     strugg
    1.04
     travellers
    0.98
     classmates
    0.93
     members
    0.85
     traveller
    0.85
     comrades
    0.81
    workers
    0.80
    Activations Density 0.095%

    No Known Activations