© 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 EvalsExportsAPI 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. 10033
    Prev
    Next
    INDEX
    Explanations

    technical instructions or code-related terms

    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
    317
    +0.13
    0.7%
    268
    +0.12
    0.6%
    358
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    317
    +0.13
    0.06
    407
    +0.12
    0.05
    219
    +0.11
    0.05
    Negative Logits
    ĻĤ
    -3.15
    ³
    -3.02
    «
    -2.83
    Ŀ
    -2.79
    µ
    -2.77
    ¾
    -2.75
    ļ
    -2.73
    ł
    -2.72
    ·
    -2.70
    °
    -2.70
    POSITIVE LOGITS
     workaround
    1.73
     argu
    1.71
     grep
    1.70
    regex
    1.65
     awk
    1.65
     regex
    1.60
    agma
    1.56
    TestCase
    1.55
    Ubuntu
    1.53
     unicode
    1.52
    Activations Density 0.917%

    No Known Activations