© 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. 22840
    Prev
    Next
    INDEX
    Explanations

    phrases that identify causality and relationships between subjects and actions

    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
    408
    +0.12
    0.5%
    647
    +0.11
    0.5%
    385
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.12
    0.05
    647
    +0.11
    0.04
    345
    +0.11
    0.04
    Negative Logits
    ©¶æ¥µ
    -0.88
    apest
    -0.87
    ensable
    -0.86
    ura
    -0.85
    ela
    -0.85
    enth
    -0.81
    imensional
    -0.81
    efer
    -0.80
    fax
    -0.80
    nai
    -0.77
    POSITIVE LOGITS
     inexper
    1.08
     inaction
    1.01
     sheer
    0.98
     deliberate
    0.95
     misunderstanding
    0.95
     incompetence
    0.93
     faulty
    0.92
     ignorance
    0.91
     lack
    0.91
     negligence
    0.91
    Activations Density 0.161%

    No Known Activations