© 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. 15144
    Prev
    Next
    INDEX
    Explanations

    expressions related to the assessment of evidence and its sufficiency or credibility

    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
    348
    +0.23
    1.3%
    78
    +0.14
    0.8%
    35
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    498
    +0.23
    0.07
    411
    +0.14
    0.06
    94
    +0.12
    0.04
    Negative Logits
    wordpress
    -1.62
    ↵³³
    -1.56
    reme
    -1.56
    LICENSE
    -1.55
     offence
    -1.54
     need
    -1.53
    æĥħ
    -1.51
     pains
    -1.51
     drama
    -1.50
     wanna
    -1.48
    POSITIVE LOGITS
    ´
    3.15
    Ń
    3.15
    Īĺ
    3.02
    ŀ
    2.98
    ¯
    2.92
     reliably
    2.89
    ²
    2.88
    ĥ½
    2.86
    ©
    2.86
    2.85
    Activations Density 0.508%

    No Known Activations