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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 1076
    Prev
    Next
    INDEX
    Explanations

    timestamps in a specific format

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    +0.14
    0.7%
    633
    +0.09
    0.4%
    501
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    403
    +0.14
    0.04
    373
    +0.09
    -0.01
    633
    +0.08
    0.04
    Negative Logits
     authority
    -0.63
     arms
    -0.62
    olis
    -0.60
     outright
    -0.59
     extermin
    -0.59
     Abrams
    -0.59
     envelope
    -0.58
     envelop
    -0.58
     pill
    -0.58
     underwear
    -0.57
    POSITIVE LOGITS
    00
    1.40
    30
    1.28
    59
    1.27
    05
    1.26
    06
    1.25
    04
    1.23
    09
    1.22
    08
    1.21
    07
    1.21
    58
    1.20
    Activations Density 0.214%

    No Known Activations