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. 242
    Prev
    Next
    INDEX
    Explanations

    social issues and discussions around rights, work, and opportunities

    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
    91
    +0.13
    0.6%
    151
    +0.11
    0.5%
    762
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    91
    +0.13
    0.18
    575
    +0.11
    0.19
    288
    +0.11
    0.22
    Negative Logits
     oneself
    -1.04
     yourselves
    -0.68
     ourselves
    -0.63
     extrad
    -0.62
    someone
    -0.61
    orks
    -0.61
    aples
    -0.60
     yourself
    -0.60
    Spawn
    -0.59
     Yourself
    -0.58
    POSITIVE LOGITS
     counterparts
    1.23
     selves
    1.16
     counterpart
    1.09
     brethren
    1.02
    mates
    0.99
    mate
    0.94
     woes
    0.93
     predecessors
    0.93
     holdings
    0.92
     arsenal
    0.92
    Activations Density 5.189%

    No Known Activations