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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 10762
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to registration and user actions in online platforms

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    271
    +0.15
    1.0%
    78
    +0.12
    0.8%
    21
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    241
    +0.15
    0.02
    59
    +0.12
    0.01
    86
    +0.11
    0.01
    Negative Logits
    soever
    -1.38
    izing
    -1.37
    merce
    -1.31
    lessly
    -1.29
    ically
    -1.25
     proportions
    -1.23
    wolf
    -1.22
    het
    -1.20
     precision
    -1.19
     negatively
    -1.19
    POSITIVE LOGITS
    àª
    1.62
    à²
    1.55
    ολ
    1.51
    ез
    1.48
    Ñij
    1.48
    еÑĤ
    1.47
    ãĥ£
    1.46
    ŀ
    1.46
    ап
    1.45
    олÑĮ
    1.44
    Activations Density 0.048%

    No Known Activations