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 24576
    5. 8-RES_FS24576-JB
    6. 631
    Prev
    Next
    INDEX
    Explanations

    the word "mean" being used in a context indicating displeasure or frustration

    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
    372
    +0.10
    0.5%
    274
    +0.10
    0.5%
    371
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    371
    +0.10
    0.02
    239
    +0.10
    0.01
    271
    +0.10
    0.01
    Negative Logits
    iets
    -0.72
    iership
    -0.66
    shaw
    -0.64
    rimp
    -0.63
    "},"
    -0.61
    è¦ļéĨĴ
    -0.60
    Desktop
    -0.59
    French
    -0.58
    appers
    -0.57
    illed
    -0.57
    POSITIVE LOGITS
     yeah
    0.84
     honestly
    0.77
     WHAT
    0.76
    entimes
    0.76
     frankly
    0.76
    terday
    0.74
     yea
    0.74
     LOOK
    0.73
     REALLY
    0.72
     obviously
    0.72
    Activations Density 0.013%

    No Known Activations