© 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. 22845
    Prev
    Next
    INDEX
    Explanations

    phrases related to supporting arguments and evidence

    oai_token-act-pair · gpt-4o-miniTriggered 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
    23
    +0.10
    0.5%
    39
    +0.10
    0.4%
    474
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.10
    0.05
    608
    +0.10
    0.05
    3
    +0.10
    0.03
    Negative Logits
     Shutterstock
    -0.72
    zie
    -0.71
    ufact
    -0.71
    soever
    -0.68
    adobe
    -0.65
    ching
    -0.62
     ILCS
    -0.60
     Screw
    -0.60
    umen
    -0.60
    lda
    -0.59
    POSITIVE LOGITS
     validity
    1.36
     assertions
    1.28
     hypotheses
    1.23
     assertion
    1.21
     hypothesis
    1.20
     authenticity
    1.14
     existence
    1.13
     claim
    1.11
     claims
    1.07
     conjecture
    1.07
    Activations Density 0.294%

    No Known Activations