© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 11328
    Prev
    Next
    INDEX
    Explanations

    terms related to luck and its effects

    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
    326
    +0.17
    0.9%
    83
    +0.13
    0.7%
    21
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.17
    0.12
    198
    +0.13
    0.10
    268
    +0.12
    0.12
    Negative Logits
    ĻĤ
    -3.58
    Ħ
    -3.04
    ¡
    -3.03
    ķ
    -2.97
    ª
    -2.88
    ī
    -2.87
    Į
    -2.80
    »
    -2.79
    ĸ
    -2.78
    ±
    -2.78
    POSITIVE LOGITS
    nier
    1.90
     capable
    1.89
    omorphic
    1.82
     incapable
    1.78
     okay
    1.71
     determinant
    1.67
     honored
    1.60
     alright
    1.60
    ogonal
    1.58
     etched
    1.58
    Activations Density 2.896%

    No Known Activations