© 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. 8966
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "a"

    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
    241
    +0.16
    1.0%
    78
    +0.15
    0.9%
    172
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    241
    +0.16
    0.03
    454
    +0.15
    0.01
    328
    +0.10
    0.01
    Negative Logits
    exports
    -1.59
     proportions
    -1.48
    jets
    -1.47
    bourg
    -1.44
     itself
    -1.44
    ois
    -1.39
     Arena
    -1.38
     bills
    -1.38
    tons
    -1.36
    aries
    -1.36
    POSITIVE LOGITS
    ī
    2.60
    Ī
    2.37
    ĭ
    2.36
    Ĵ
    2.36
    ffected
    2.20
    ľ
    2.17
    ĵ
    2.08
    ¡
    2.07
    ¸
    2.02
    ª
    2.01
    Activations Density 0.052%

    No Known Activations