© 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. 3014
    Prev
    Next
    INDEX
    Explanations

    direct speech quotations

    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
    607
    +0.09
    0.4%
    604
    +0.09
    0.4%
    300
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    100
    +0.09
    0.08
    607
    +0.09
    0.07
    105
    +0.09
    0.07
    Negative Logits
     occasion
    -0.69
     Flavoring
    -0.66
     subcommittee
    -0.65
     surfaces
    -0.63
    iffe
    -0.61
     princip
    -0.61
     pole
    -0.60
    imore
    -0.59
    unal
    -0.59
     Brune
    -0.59
    POSITIVE LOGITS
    Hey
    1.30
     hey
    1.07
    hey
    1.05
    hello
    0.99
    Dear
    0.99
    Fuck
    0.98
    Oh
    0.98
     Hey
    0.98
    why
    0.96
    Hello
    0.95
    Activations Density 1.431%

    No Known Activations