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. 1131
    Prev
    Next
    INDEX
    Explanations

    background information or context within a discussion, often related to personal experiences or observations

    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
    7
    +0.12
    0.5%
    126
    +0.11
    0.5%
    395
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    395
    +0.12
    0.09
    126
    +0.11
    0.09
    408
    +0.11
    0.11
    Negative Logits
    assisted
    -0.77
    OWS
    -0.70
    ares
    -0.70
    buster
    -0.70
    ãĥ¼ãĤ¯
    -0.66
    his
    -0.66
    friends
    -0.66
    hammad
    -0.64
    ilk
    -0.64
    Results
    -0.63
    POSITIVE LOGITS
     overlap
    1.04
     difference
    1.02
     shortage
    1.02
     possibility
    0.98
     inherent
    0.97
     reason
    0.95
     discrepancy
    0.94
     misconception
    0.93
     mismatch
    0.92
     lurking
    0.92
    Activations Density 1.298%

    No Known Activations