Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 233
    Prev
    Next
    INDEX
    Explanations

    specific technical terms related to software development and programming

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    377
    +0.15
    0.7%
    365
    +0.13
    0.6%
    549
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.15
    0.23
    365
    +0.13
    0.07
    659
    +0.12
    0.13
    Negative Logits
    vous
    -0.80
    ingen
    -0.68
    soDeliveryDate
    -0.68
    stakes
    -0.66
    risome
    -0.64
    hops
    -0.63
     Fargo
    -0.61
    士
    -0.60
     SEAL
    -0.60
    tty
    -0.59
    POSITIVE LOGITS
    illary
    1.09
    osite
    0.90
    venture
    0.90
    qua
    0.89
    alys
    0.84
    alon
    0.83
    iom
    0.82
    osta
    0.81
    raham
    0.81
    reement
    0.79
    Activations Density 5.605%

    No Known Activations