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

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 172
    Prev
    Next
    INDEX
    Explanations

    information related to events and articles on a website

    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
    87
    +0.39
    2.7%
    756
    +0.27
    1.9%
    459
    +0.24
    1.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.39
    -0.13
    756
    +0.27
    -0.09
    459
    +0.24
    0.32
    Negative Logits
     Canaver
    -0.57
    iannopoulos
    -0.45
    ¶
    -0.45
    sect
    -0.39
    RESULTS
    -0.39
    osate
    -0.38
    ãĥ¤
    -0.38
    POST
    -0.36
    ?????-?????-
    -0.36
     Yiannopoulos
    -0.36
    POSITIVE LOGITS
     however
    0.38
     substituted
    0.38
     latter
    0.37
     afterwards
    0.36
     likewise
    0.36
     displ
    0.35
     vou
    0.35
    itto
    0.35
     replaced
    0.34
     whereas
    0.34
    Activations Density 24.509%

    No Known Activations