© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 1536
    9. 8-RES_FS1536-JB
    11. 80
    Prev
    Next
    INDEX
    Explanations

    mentions of social media accounts, specifically Twitter

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    13
    +0.11
    0.5%
    294
    +0.11
    0.5%
    396
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    396
    +0.11
    0.07
    13
    +0.11
    0.09
    294
    +0.10
    0.06
    Negative Logits
    cised
    -0.75
    ellectual
    -0.67
     CFR
    -0.66
    cephal
    -0.64
    minded
    -0.62
     Tripoli
    -0.62
     circumcised
    -0.62
     Croat
    -0.62
    urion
    -0.61
     Quart
    -0.60
    POSITIVE LOGITS
    username
    1.02
     Username
    1.01
     username
    0.96
     hashtag
    0.93
     (@
    0.92
     feeds
    0.90
     "@
    0.89
     hasht
    0.88
     postings
    0.86
     accounts
    0.83
    Activations Density 1.553%

    No Known Activations