© 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 EvalsExportsinterp-engineNEWAPI 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. 27811
    Prev
    Next
    INDEX
    Explanations

    technical programming or coding syntax elements

    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
    233
    +0.14
    0.8%
    146
    +0.12
    0.7%
    400
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.14
    0.12
    160
    +0.12
    0.14
    130
    +0.12
    0.16
    Negative Logits
    ±
    -2.40
    §
    -2.33
    Ĵ
    -2.25
    -2.17
    ↵↵
    -2.17
    -2.17
    -2.17
    ↵
    -2.17
    ↵
    -2.17
    ↵↵↵
    -2.17
    POSITIVE LOGITS
     recess
    1.61
     unlike
    1.55
     lately
    1.45
     pyl
    1.42
     whatsoever
    1.40
    ipes
    1.36
    soever
    1.36
     because
    1.35
     someday
    1.35
     defeating
    1.34
    Activations Density 5.246%

    No Known Activations