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 768
    5. 8-RES_FS768-JB
    6. 507
    Prev
    Next
    INDEX
    Explanations

    phrases related to security, investigation, and government actions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    480
    +0.66
    8.8%
    266
    +0.15
    2.0%
    420
    +0.06
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.66
    0.07
    266
    +0.15
    0.43
    284
    +0.06
    0.10
    Negative Logits
    anwhile
    -0.20
    abee
    -0.18
     furthermore
    -0.17
     compliant
    -0.17
     meanwhile
    -0.17
    nesday
    -0.17
     onward
    -0.17
     vou
    -0.17
    ardless
    -0.17
     moreover
    -0.17
    POSITIVE LOGITS
    natureconservancy
    0.23
    cffffcc
    0.21
    taboola
    0.20
     âĢº
    0.20
    )--
    0.18
    )=(
    0.18
    à¨
    0.17
    catentry
    0.17
    acterial
    0.16
    ãĥ»
    0.16
    Activations Density 21.737%

    No Known Activations