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

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 2807
    Prev
    Next
    INDEX
    Explanations

    phrases related to news and technology advancements

    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
    371
    +0.13
    0.6%
    506
    +0.12
    0.5%
    679
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    506
    +0.13
    0.11
    371
    +0.12
    0.12
    204
    +0.10
    0.09
    Negative Logits
    fw
    -0.66
     [];
    -0.66
    MpServer
    -0.65
    imester
    -0.64
    Request
    -0.63
    word
    -0.63
     snowball
    -0.62
    perate
    -0.62
    aber
    -0.61
    wed
    -0.60
    POSITIVE LOGITS
     Shaun
    1.32
     extraord
    1.26
     Stephen
    1.24
     Richard
    1.24
     Erik
    1.23
     Tony
    1.22
     Larry
    1.21
     Michael
    1.21
     Andrew
    1.20
     Kevin
    1.20
    Activations Density 1.030%

    No Known Activations