© 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 3072
    9. 8-RES_FS3072-JB
    11. 3021
    Prev
    Next
    INDEX
    Explanations

    keywords related to politics and legislation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    380
    +0.13
    0.6%
    721
    +0.11
    0.5%
    192
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    380
    +0.13
    0.07
    43
    +0.11
    0.07
    59
    +0.09
    0.07
    Negative Logits
    ongevity
    -0.70
    Shut
    -0.70
    yt
    -0.70
    role
    -0.69
    Cover
    -0.65
    Throw
    -0.65
    enz
    -0.65
     </
    -0.64
    EngineDebug
    -0.64
    arat
    -0.64
    POSITIVE LOGITS
     however
    0.97
     researchers
    0.78
     meanwhile
    0.78
     analysts
    0.72
     we
    0.72
     assailants
    0.69
     which
    0.68
     staffers
    0.67
     Gina
    0.66
     researcher
    0.66
    Activations Density 0.902%

    No Known Activations