© 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. 30768
    Prev
    Next
    INDEX
    Explanations

    specific symbols or formatting elements in structured text

    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
    156
    +0.60
    4.7%
    111
    +0.45
    3.6%
    271
    +0.14
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.60
    0.23
    111
    +0.45
    0.24
    71
    +0.14
    0.20
    Negative Logits
    hores
    -1.47
    uated
    -1.30
    uates
    -1.26
    akes
    -1.23
    ites
    -1.18
    uate
    -1.13
    ake
    -1.11
    asone
    -1.07
    aked
    -1.03
     \[
    -1.03
    POSITIVE LOGITS
    "}](#
    1.73
    neum
    1.25
    »¿
    1.22
    ligt
    1.18
     bang
    1.17
    amsfonts
    1.17
    į
    1.14
    edly
    1.06
    ^](#
    1.03
     pyl
    1.03
    Activations Density 0.229%

    No Known Activations