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. 280
    Prev
    Next
    INDEX
    Explanations

    persons responsible for criminal or violent acts

    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
    66
    +0.10
    0.4%
    91
    +0.10
    0.4%
    447
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    247
    +0.10
    0.10
    91
    +0.10
    0.13
    288
    +0.09
    0.15
    Negative Logits
    ories
    -0.81
    izoph
    -0.67
    ilion
    -0.66
    earchers
    -0.66
    etting
    -0.65
    aeda
    -0.65
    equality
    -0.65
     Invalid
    -0.65
    Bridge
    -0.65
    arus
    -0.63
    POSITIVE LOGITS
     who
    1.28
     whom
    1.14
    who
    1.11
    hood
    0.99
     named
    0.95
     surn
    0.93
     whose
    0.89
    's
    0.88
    esses
    0.85
     WHO
    0.85
    Activations Density 2.636%

    No Known Activations