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 3072
    5. 8-RES_FS3072-JB
    6. 1573
    Prev
    Next
    INDEX
    Explanations

    benefits or advantages related to a given topic or situation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    420
    +0.14
    0.6%
    285
    +0.12
    0.6%
    288
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.09
    285
    +0.12
    0.05
    444
    +0.12
    0.06
    Negative Logits
     Gears
    -0.67
    metal
    -0.64
     Adin
    -0.64
     Penal
    -0.64
    rez
    -0.61
    opic
    -0.59
    ãĥ£
    -0.59
    netflix
    -0.59
     Milky
    -0.57
     Alma
    -0.57
    POSITIVE LOGITS
     conferred
    1.16
     outwe
    1.08
     outweigh
    0.98
     afforded
    0.98
    itism
    0.96
     bestowed
    0.92
    Reviewer
    0.86
    ously
    0.82
    benefit
    0.82
     accrued
    0.81
    Activations Density 0.962%

    No Known Activations