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 768
    5. 8-RES_FS768-JB
    6. 221
    Prev
    Next
    INDEX
    Explanations

    phrases containing the words 'no time', 'any way', 'season', 'primaries', 'others', 'Tor', 'virtual', 'game', 'force', 'comp', 'et', 'turning', 'Flint', 'statement', 'doubt', 'House',

    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
    355
    +0.13
    0.6%
    277
    +0.10
    0.5%
    599
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    355
    +0.13
    0.13
    599
    +0.10
    0.09
    402
    +0.10
    0.10
    Negative Logits
    apego
    -0.73
    inx
    -0.65
    ulner
    -0.64
    itiz
    -0.61
    ourge
    -0.60
    letes
    -0.60
    orsi
    -0.59
    paio
    -0.59
    ographers
    -0.59
     Yorkers
    -0.58
    POSITIVE LOGITS
     guise
    1.28
     manner
    1.20
     vicinity
    1.15
     context
    1.13
     contexts
    1.09
     nutshell
    1.09
     fashion
    1.05
     midst
    1.02
     circles
    1.02
     vein
    1.01
    Activations Density 4.960%

    No Known Activations