© 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. 2-RES-SM
    11. 26323
    Prev
    Next
    INDEX
    Explanations

    questions and words indicating uncertainty or inquiry

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    419
    +0.16
    0.9%
    287
    +0.13
    0.7%
    320
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    44
    +0.16
    0.06
    395
    +0.13
    0.07
    284
    +0.12
    0.06
    Negative Logits
    ĻĤ
    -5.26
    Ļª
    -5.08
    Īĺ
    -4.96
    »¿
    -4.71
    ©
    -4.68
    Ĵ
    -4.61
    ĺ
    -4.48
    ¸
    -4.46
    ŀ
    -4.44
    «
    -4.43
    POSITIVE LOGITS
    others
    2.03
     consequent
    1.95
     others
    1.91
     thereof
    1.69
    other
    1.68
     accompanying
    1.60
     omission
    1.52
    anges
    1.49
     importantly
    1.49
    therefore
    1.48
    Activations Density 0.586%

    No Known Activations