© 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. 9741
    Prev
    Next
    INDEX
    Explanations

    technical terms related to software or coding

    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
    481
    +0.19
    0.9%
    447
    +0.18
    0.9%
    314
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    314
    +0.19
    0.01
    326
    +0.18
    0.06
    701
    +0.14
    0.04
    Negative Logits
     dismantling
    -0.73
     scraps
    -0.66
     accountability
    -0.65
     experiments
    -0.65
     adherence
    -0.64
     wallets
    -0.63
     incentives
    -0.62
     implementations
    -0.62
     legacy
    -0.62
     archives
    -0.61
    POSITIVE LOGITS
    -.
    1.09
    /,
    1.02
    É
    0.98
    Äĵ
    0.94
    _.
    0.92
    -,
    0.90
    qq
    0.87
    foo
    0.86
    XXX
    0.84
    Åį
    0.82
    Activations Density 0.550%

    No Known Activations