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 1536
    5. 8-RES_FS1536-JB
    6. 1174
    Prev
    Next
    INDEX
    Explanations

    medical conditions and procedures

    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
    373
    +0.17
    0.8%
    640
    +0.16
    0.7%
    447
    +0.14
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.17
    0.17
    160
    +0.16
    0.11
    373
    +0.14
    -0.04
    Negative Logits
    hement
    -0.75
     HOU
    -0.65
    plet
    -0.63
     repro
    -0.60
     commits
    -0.60
     wet
    -0.59
     rolls
    -0.57
    ivable
    -0.57
     taps
    -0.57
     DOT
    -0.56
    POSITIVE LOGITS
    ccording
    1.21
    Spoiler
    0.82
    Throughout
    0.80
    Sometimes
    0.79
     Firstly
    0.78
     Initially
    0.76
     Originally
    0.76
    Alright
    0.74
    Firstly
    0.74
     Throughout
    0.73
    Activations Density 1.256%

    No Known Activations