© 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. 10184
    Prev
    Next
    INDEX
    Explanations

    technical terms related to electronic gadgets and hardware

    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
    640
    +0.35
    1.7%
    447
    +0.16
    0.8%
    496
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.35
    0.16
    6
    +0.16
    0.12
    107
    +0.12
    0.07
    Negative Logits
     freel
    -0.88
     divest
    -0.84
     defund
    -0.80
     arguing
    -0.80
     volunt
    -0.79
     invoking
    -0.77
     altru
    -0.76
     quitting
    -0.76
     attrition
    -0.76
     undecided
    -0.75
    POSITIVE LOGITS
    Features
    1.42
    Includes
    1.42
    Ingredients
    1.26
     Specifications
    1.25
    Optional
    1.24
    Mini
    1.24
    Rated
    1.23
    Requires
    1.23
     Includes
    1.22
    Available
    1.21
    Activations Density 0.617%

    No Known Activations