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 24576
    5. 8-RES_FS24576-JB
    6. 18975
    Prev
    Next
    INDEX
    Explanations

    phrases related to politics and constituents

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    87
    +0.42
    2.3%
    255
    +0.11
    0.6%
    481
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.42
    -0.01
    762
    +0.11
    0.10
    756
    +0.10
    -0.02
    Negative Logits
    SourceFile
    -0.57
     Released
    -0.56
    MpServer
    -0.55
     Orchestra
    -0.53
     confir
    -0.53
    acas
    -0.51
    yssey
    -0.51
    Lua
    -0.50
    las
    -0.50
    itled
    -0.48
    POSITIVE LOGITS
     rather
    0.89
     anyway
    0.88
     wherever
    0.86
     everywhere
    0.85
     anyways
    0.79
     insofar
    0.77
     oneself
    0.75
     blindly
    0.73
     endlessly
    0.72
     inherently
    0.72
    Activations Density 1.031%

    No Known Activations