© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 7374
    Prev
    Next
    INDEX
    Explanations

    non-English characters and phrases from various languages, such as Armenian, Urdu, Arabic, Persian, Kurdish, Marathi, Hindi, and Icelandic

    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
    48
    +0.12
    0.6%
    256
    +0.10
    0.4%
    213
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    103
    +0.12
    0.01
    107
    +0.10
    0.01
    48
    +0.10
    0.01
    Negative Logits
    ettings
    -0.77
    aturdays
    -0.73
    ovie
    -0.70
    essa
    -0.65
     Compton
    -0.65
     Osc
    -0.64
    sonian
    -0.62
     Ness
    -0.61
    estern
    -0.61
     Rossi
    -0.60
    POSITIVE LOGITS
    ¤
    1.58
    ¹
    1.49
    Ĥ
    1.47
    ¸
    1.45
    ¦
    1.44
    ®
    1.42
    ²
    1.42
    ¨
    1.41
    ļ
    1.40
    °
    1.39
    Activations Density 0.007%

    No Known Activations