© 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. 31005
    Prev
    Next
    INDEX
    Explanations

    XML attributes and their values

    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
    479
    +0.12
    0.7%
    203
    +0.11
    0.6%
    47
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    385
    +0.12
    0.02
    165
    +0.11
    0.02
    74
    +0.11
    0.02
    Negative Logits
    plantation
    -1.52
    ectomy
    -1.50
     runway
    -1.47
     effect
    -1.46
     opportunity
    -1.46
     diversion
    -1.44
    oret
    -1.43
     winning
    -1.41
     receives
    -1.41
    owe
    -1.39
    POSITIVE LOGITS
    °
    2.85
    ¸
    2.84
    Īĺ
    2.81
    İ
    2.75
    Ģ
    2.75
    ¯
    2.72
    ª
    2.72
    ¥
    2.69
    ´
    2.69
    ³
    2.67
    Activations Density 0.087%

    No Known Activations