© 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. 26195
    Prev
    Next
    INDEX
    Explanations

    references to skills and abilities relevant to job qualifications

    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
    274
    +0.15
    0.8%
    141
    +0.11
    0.7%
    140
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    140
    +0.15
    0.02
    141
    +0.11
    0.02
    116
    +0.11
    0.02
    Negative Logits
    ·¸
    -3.73
    ³
    -3.61
    °
    -3.37
    Ń
    -3.35
    µ
    -3.23
    ĻĤ
    -3.23
    ·
    -3.14
    Ģ
    -3.13
    ´
    -3.11
    «
    -3.09
    POSITIVE LOGITS
    ful
    1.98
     skills
    1.86
     level
    1.73
     Skills
    1.71
    omed
    1.70
    management
    1.66
    balanced
    1.64
    less
    1.60
    metry
    1.55
    ually
    1.53
    Activations Density 0.229%

    No Known Activations