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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 43417
    Prev
    Next
    INDEX
    Explanations

    conditional phrases indicating consequences or alternatives

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    485
    +0.10
    0.4%
    248
    +0.09
    0.4%
    596
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    248
    +0.10
    0.01
    8
    +0.09
    0.01
    133
    +0.09
    0.01
    Negative Logits
    levard
    -0.65
    foot
    -0.60
    è¦ļéĨĴ
    -0.59
    orig
    -0.56
    nesday
    -0.56
    uto
    -0.55
    zon
    -0.55
    ste
    -0.54
    stration
    -0.53
    aez
    -0.53
    POSITIVE LOGITS
    SPONSORED
    0.70
    annon
    0.67
     Tactics
    0.64
     proceed
    0.63
    wise
    0.63
    thodox
    0.63
    GAN
    0.62
    ally
    0.62
    forth
    0.60
     Proceed
    0.60
    Activations Density 0.012%

    No Known Activations