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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 11183
    Prev
    Next
    INDEX
    Explanations

    JavaScript function code snippets for adding scripts and working with memory addresses

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    83
    +0.11
    0.5%
    640
    +0.11
    0.5%
    694
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    257
    +0.11
    0.02
    83
    +0.11
    0.02
    107
    +0.09
    0.02
    Negative Logits
     elusive
    -0.72
     GOODMAN
    -0.71
     trance
    -0.63
     hackers
    -0.62
     journalism
    -0.61
     gamers
    -0.60
     bandits
    -0.60
    ucha
    -0.60
     pir
    -0.60
     lurking
    -0.60
    POSITIVE LOGITS
    append
    1.02
    mk
    1.01
    remove
    0.99
    create
    0.96
    apply
    0.96
    prototype
    0.96
    get
    0.96
    name
    0.94
    length
    0.94
    find
    0.94
    Activations Density 0.039%

    No Known Activations