© 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. 31877
    Prev
    Next
    INDEX
    Explanations

    numerical information and coding-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
    263
    +0.16
    1.0%
    210
    +0.14
    0.8%
    217
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    469
    +0.16
    0.09
    59
    +0.14
    0.04
    121
    +0.13
    0.07
    Negative Logits
     niche
    -1.53
    oretic
    -1.51
    istered
    -1.45
     differences
    -1.42
    ural
    -1.40
    ylan
    -1.39
     difference
    -1.34
    ões
    -1.29
    vid
    -1.27
    openh
    -1.25
    POSITIVE LOGITS
    ¿½
    1.77
    ĻĤ
    1.74
    NESS
    1.54
    ¼
    1.42
    ¦
    1.41
    ĺ
    1.29
    ERN
    1.26
     thereafter
    1.26
     territory
    1.26
     hose
    1.24
    Activations Density 4.512%

    No Known Activations