© 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 EvalsExportsAPI 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. 18248
    Prev
    Next
    INDEX
    Explanations

    verbs with an infinitive 'to' followed by another verb, such as 'to walk', 'to stand', and 'to make'

    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
    665
    +0.13
    0.6%
    608
    +0.12
    0.6%
    97
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.13
    0.03
    718
    +0.12
    0.03
    49
    +0.11
    0.03
    Negative Logits
    lights
    -0.88
    teen
    -0.77
    listed
    -0.76
    holder
    -0.70
     Appears
    -0.70
    isite
    -0.67
     Hartford
    -0.66
    termination
    -0.65
    prem
    -0.63
     Coff
    -0.63
    POSITIVE LOGITS
     behave
    1.25
     differentiate
    1.23
     cope
    1.23
     recognize
    1.20
     tolerate
    1.15
     manipulate
    1.14
     wrestle
    1.12
     distinguish
    1.09
     imitate
    1.07
     recognise
    1.06
    Activations Density 0.073%

    No Known Activations