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

    Neuronpedia

    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. 2-RES-SM
    11. 30011
    Prev
    Next
    INDEX
    Explanations

    the word "the"

    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
    98
    +0.15
    0.8%
    394
    +0.12
    0.7%
    335
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    394
    +0.15
    0.02
    335
    +0.12
    0.02
    245
    +0.11
    0.02
    Negative Logits
     Parenthood
    -1.51
    ules
    -1.49
    aign
    -1.42
    \](
    -1.38
    emas
    -1.38
     homem
    -1.36
     figures
    -1.30
    esque
    -1.30
     smokers
    -1.29
     succ
    -1.29
    POSITIVE LOGITS
     *)
    1.69
     underlying
    1.61
    NdEx
    1.57
     GNU
    1.49
    following
    1.46
     predecessor
    1.45
     following
    1.43
    rieved
    1.40
     transcript
    1.35
    ὶ
    1.33
    Activations Density 0.042%

    No Known Activations