Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 467
    Prev
    Next
    INDEX
    Explanations

    political and economic themes related to government interventions, economic conditions, and ideological beliefs

    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
    321
    +0.11
    0.5%
    238
    +0.11
    0.5%
    489
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.11
    0.13
    36
    +0.11
    0.12
    489
    +0.09
    0.09
    Negative Logits
    upon
    -0.77
    ritch
    -0.70
    ocate
    -0.67
    atron
    -0.67
    ibur
    -0.63
    ties
    -0.63
    ̶
    -0.62
    orically
    -0.59
    lance
    -0.59
    lessly
    -0.58
    POSITIVE LOGITS
     occasions
    1.48
     basis
    1.39
     occasion
    1.33
     behalf
    1.31
     doorstep
    1.28
     grounds
    1.18
     sidelines
    1.18
     verge
    1.17
     eve
    1.14
     heels
    1.10
    Activations Density 2.299%

    No Known Activations