© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 27852
    Prev
    Next
    INDEX
    Explanations

    references to single items or singular concepts

    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
    282
    +0.14
    0.8%
    156
    +0.13
    0.7%
    436
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    249
    +0.14
    0.02
    101
    +0.13
    0.01
    196
    +0.13
    0.02
    Negative Logits
    ses
    -1.60
    frastructure
    -1.60
    rich
    -1.50
    ruvate
    -1.50
    owed
    -1.47
    inherit
    -1.45
    angers
    -1.45
    olesterol
    -1.44
    rypt
    -1.41
    alls
    -1.41
    POSITIVE LOGITS
    meal
    1.81
     digits
    1.79
    ¿
    1.76
    Īĺ
    1.75
    į
    1.72
    shot
    1.53
     sized
    1.50
    eling
    1.48
     handful
    1.47
    berg
    1.46
    Activations Density 0.059%

    No Known Activations