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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 11437
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "be" and related concepts of existence or being

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    267
    +0.13
    0.8%
    419
    +0.13
    0.7%
    107
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    267
    +0.13
    0.02
    107
    +0.13
    0.01
    359
    +0.12
    0.01
    Negative Logits
    \'
    -2.02
    ment
    -1.79
    tion
    -1.77
    ÃŃs
    -1.64
    âĢIJ
    -1.64
    might
    -1.62
    ...](
    -1.58
    may
    -1.57
    RUPT
    -1.55
    like
    -1.54
    POSITIVE LOGITS
     wk
    1.55
     {{
    1.47
     dz
    1.44
     Appeal
    1.43
     \|_{
    1.41
     [[
    1.40
    dorff
    1.39
     nil
    1.36
    "|
    1.35
    Ļª
    1.35
    Activations Density 0.035%

    No Known Activations