© 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. 30368
    Prev
    Next
    INDEX
    Explanations

    occurrences of numerical data and programming error messages

    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
    410
    +0.16
    0.9%
    126
    +0.10
    0.6%
    156
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    209
    +0.16
    0.07
    410
    +0.10
    0.08
    291
    +0.10
    0.06
    Negative Logits
    ourse
    -1.79
    ould
    -1.71
    uchi
    -1.65
    stone
    -1.58
    ?’
    -1.53
    com
    -1.48
    abil
    -1.44
    ados
    -1.44
    ashi
    -1.43
    hl
    -1.37
    POSITIVE LOGITS
     heartbeat
    1.69
     synd
    1.51
    ourcing
    1.50
     Minister
    1.47
     briefing
    1.46
    //
    1.46
     minister
    1.46
     swallow
    1.45
     echo
    1.36
     hers
    1.35
    Activations Density 0.379%

    No Known Activations