© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 2601
    Prev
    Next
    INDEX
    Explanations

    references to actions requiring response or interaction, possibly related to tech security breaches

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    +0.27
    1.3%
    640
    +0.21
    1.0%
    160
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.27
    0.06
    314
    +0.21
    0.02
    640
    +0.12
    0.05
    Negative Logits
     prey
    -0.70
     drift
    -0.70
     imperson
    -0.69
     occupancy
    -0.68
     clin
    -0.68
     flee
    -0.67
     extingu
    -0.67
     electr
    -0.66
     spir
    -0.66
     depreciation
    -0.66
    POSITIVE LOGITS
    PDF
    1.49
    emphasis
    1.32
    see
    1.31
    via
    1.31
    pdf
    1.30
    below
    1.27
    please
    1.25
    sorry
    1.22
    albeit
    1.21
    pictured
    1.20
    Activations Density 0.097%

    No Known Activations