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 1536
    5. 8-RES_FS1536-JB
    6. 886
    Prev
    Next
    INDEX
    Explanations

    phrases related to events or detailed descriptions

    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
    724
    +0.11
    0.5%
    581
    +0.11
    0.5%
    758
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    159
    +0.11
    0.10
    299
    +0.11
    0.09
    758
    +0.10
    0.06
    Negative Logits
    gall
    -0.73
    venants
    -0.69
    cheat
    -0.68
    Cho
    -0.68
    anon
    -0.65
    ibr
    -0.61
    etheless
    -0.61
    bek
    -0.58
     Cheong
    -0.57
    Lock
    -0.57
    POSITIVE LOGITS
     sake
    2.20
     purposes
    1.99
     reasons
    1.39
    ummies
    1.34
     purpose
    1.18
     foreseeable
    1.05
     reason
    0.98
     duration
    0.92
     Reasons
    0.90
     upcoming
    0.88
    Activations Density 2.747%

    No Known Activations