© 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. 4-RES-SM
    11. 11061
    Prev
    Next
    INDEX
    Explanations

    the word "even" 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/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
    243
    +0.12
    0.7%
    207
    +0.12
    0.7%
    139
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    207
    +0.12
    0.03
    243
    +0.12
    0.05
    402
    +0.11
    0.05
    Negative Logits
    atial
    -1.60
    tract
    -1.58
    ĻĤ
    -1.56
    ħ
    -1.48
    ¿
    -1.43
    ract
    -1.40
    /\
    -1.39
    ellar
    -1.39
    ient
    -1.38
    Ģ
    -1.37
    POSITIVE LOGITS
     though
    2.85
    handed
    2.73
     worse
    2.33
     remotely
    2.07
     if
    2.06
     tho
    1.97
     stranger
    1.85
    though
    1.82
    ably
    1.78
     tougher
    1.77
    Activations Density 0.153%

    No Known Activations