© 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 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. 30475
    Prev
    Next
    INDEX
    Explanations

    the frequency of the word "well" used in various contexts

    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
    199
    +0.12
    0.6%
    307
    +0.11
    0.6%
    184
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    199
    +0.12
    0.02
    477
    +0.11
    0.04
    184
    +0.10
    0.02
    Negative Logits
    Į
    -1.77
    ļ
    -1.77
    ribe
    -1.76
    FROM
    -1.60
    atar
    -1.59
    ŀ
    -1.54
    Ļ
    -1.52
    Decided
    -1.51
    Ŀ
    -1.49
    ¯
    -1.47
    POSITIVE LOGITS
    bred
    2.16
    ified
    1.69
     underway
    1.68
     suited
    1.66
     behaved
    1.66
     regarded
    1.63
    icht
    1.54
    spring
    1.52
     positioned
    1.50
    founded
    1.50
    Activations Density 0.110%

    No Known Activations