© 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. 10290
    Prev
    Next
    INDEX
    Explanations

    instances of the word "rest" to signify relaxation or recovery

    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
    320
    +0.12
    0.7%
    115
    +0.12
    0.7%
    125
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    389
    +0.12
    0.03
    251
    +0.12
    0.02
    115
    +0.12
    0.02
    Negative Logits
    ĵ
    -2.77
    ¾
    -2.43
    ľĵ
    -2.42
    ¶
    -2.26
    ĸ´
    -2.25
    ľ
    -2.21
    ¿
    -2.20
    ĨĴ
    -2.15
    ĻĤ
    -2.15
    ¯
    -2.12
    POSITIVE LOGITS
    lessness
    2.38
    aurant
    2.20
    lessly
    2.15
    ructure
    2.03
    raints
    1.91
    atement
    1.83
    ful
    1.82
    rict
    1.76
    rooms
    1.75
    riction
    1.74
    Activations Density 0.190%

    No Known Activations