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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 936
    Prev
    Next
    INDEX
    Explanations

    technical terms related to chemical procedures and laboratory techniques

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    471
    +0.12
    0.5%
    234
    +0.10
    0.5%
    167
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    230
    +0.12
    0.08
    310
    +0.10
    0.08
    471
    +0.10
    0.07
    Negative Logits
    achi
    -0.71
    abol
    -0.65
    iqu
    -0.65
    alyses
    -0.65
    warts
    -0.63
     unh
    -0.61
     unsc
    -0.61
    oles
    -0.61
    elly
    -0.61
    verend
    -0.61
    POSITIVE LOGITS
    Hz
    1.16
    px
    1.11
     MHz
    1.06
    MHz
    1.03
    +
    1.00
    Mbps
    0.98
    mAh
    0.98
    00000
    0.96
    hz
    0.95
     Hz
    0.95
    Activations Density 0.467%

    No Known Activations