© 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. 12590
    Prev
    Next
    INDEX
    Explanations

    programming-related errors and exceptions

    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
    121
    +0.17
    1.0%
    31
    +0.15
    0.9%
    198
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    121
    +0.17
    0.09
    360
    +0.15
    0.07
    400
    +0.14
    0.02
    Negative Logits
     pockets
    -1.63
    ew
    -1.56
     Workers
    -1.46
     archives
    -1.46
     Walmart
    -1.45
     ),
    -1.43
     alumni
    -1.40
     shelves
    -1.39
     Archives
    -1.38
    artments
    -1.38
    POSITIVE LOGITS
    ²
    2.39
    ĥ½
    2.07
    Ĥ¬
    2.01
    ĻĤ
    2.01
    Ħ
    1.93
    ĨĴ
    1.84
    <|outofrange|>
    1.82
    ↵↵↵
    1.82
    ↵
    1.82
    1.82
    Activations Density 6.785%

    No Known Activations