© 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. 1436
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal and policy matters

    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
    314
    +0.17
    0.8%
    160
    +0.12
    0.5%
    218
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    152
    +0.17
    0.07
    314
    +0.12
    0.03
    124
    +0.11
    0.06
    Negative Logits
     beads
    -0.71
     Bullets
    -0.67
     Dunk
    -0.66
     Rounds
    -0.65
     Downs
    -0.65
     Forbes
    -0.65
     sunset
    -0.63
     Cotton
    -0.62
     ceremony
    -0.62
     buck
    -0.62
    POSITIVE LOGITS
    centric
    1.31
    friendly
    1.25
    centered
    1.24
    specific
    1.19
    oriented
    1.18
    driven
    1.16
    controlled
    1.16
    focused
    1.12
    induced
    1.10
    generated
    1.08
    Activations Density 0.136%

    No Known Activations