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 24576
    5. 8-RES_FS24576-JB
    6. 23655
    Prev
    Next
    INDEX
    Explanations

    the word "as" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    314
    +0.10
    0.5%
    320
    +0.10
    0.4%
    168
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    192
    +0.10
    0.04
    441
    +0.10
    0.04
    147
    +0.10
    0.04
    Negative Logits
    æĸ
    -0.64
    ries
    -0.59
    roads
    -0.58
    rior
    -0.57
    artifacts
    -0.57
    aster
    -0.55
    agos
    -0.55
    Sharp
    -0.54
    rous
    -0.54
    Psy
    -0.54
    POSITIVE LOGITS
    pired
    1.22
    pires
    1.15
    phalt
    1.11
    pects
    1.06
    bestos
    1.06
    cription
    1.04
    semb
    1.02
    pire
    0.99
    pect
    0.99
    semble
    0.98
    Activations Density 0.115%

    No Known Activations