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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 818
    Prev
    Next
    INDEX
    Explanations

    information related to rules, regulations, requirements, and criteria

    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
    21
    +0.10
    0.5%
    605
    +0.10
    0.5%
    46
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.10
    0.05
    718
    +0.10
    0.04
    738
    +0.10
    0.03
    Negative Logits
    blogspot
    -0.82
    Written
    -0.77
    TL
    -0.70
     contrace
    -0.69
    cycl
    -0.64
    wordpress
    -0.63
    dozen
    -0.61
     drifted
    -0.59
    LESS
    -0.59
    kin
    -0.56
    POSITIVE LOGITS
     qualify
    1.22
     justify
    0.95
     survive
    0.91
     qualifies
    0.86
     eligible
    0.86
     gain
    0.85
     achieve
    0.85
     earn
    0.85
     guarantee
    0.84
     access
    0.84
    Activations Density 0.193%

    No Known Activations