© 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. 4-RES-SM
    11. 14276
    Prev
    Next
    INDEX
    Explanations

    words related to gluten and its health effects

    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
    +0.17
    1.0%
    394
    +0.11
    0.6%
    108
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    394
    +0.17
    0.06
    486
    +0.11
    0.05
    108
    +0.11
    0.05
    Negative Logits
    chaft
    -1.56
    ships
    -1.54
     behalf
    -1.53
    iciary
    -1.47
    encephal
    -1.46
    pace
    -1.46
    orov
    -1.45
    estly
    -1.43
     themselves
    -1.39
    icolor
    -1.39
    POSITIVE LOGITS
    ategories
    2.33
    andidate
    2.27
    esar
    2.19
    aused
    2.10
    Ĵ
    2.07
    ursor
    2.05
    rystall
    2.04
    offe
    2.00
    ORE
    1.99
     elegans
    1.98
    Activations Density 0.071%

    No Known Activations