© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 15331
    Prev
    Next
    INDEX
    Explanations

    adjectives and terms related to weight and body image

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.36
    2.2%
    25
    +0.16
    1.0%
    9
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.36
    0.06
    401
    +0.16
    0.02
    509
    +0.15
    0.08
    Negative Logits
    ¸
    -2.32
    ĻĤ
    -2.24
    ij
    -2.21
    Ŀ
    -2.19
    ¹
    -2.07
    ½
    -2.06
    Ĭ
    -2.00
    Ľ
    -1.99
    ĺ
    -1.97
    ¤
    -1.96
    POSITIVE LOGITS
    ting
    3.15
    ty
    2.99
    ters
    2.69
    ly
    2.63
    ial
    2.51
    iation
    2.50
    ially
    2.46
    ional
    2.43
    ment
    2.38
    less
    2.36
    Activations Density 2.329%

    No Known Activations