© 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. 4-RES-SM
    11. 14721
    Prev
    Next
    INDEX
    Explanations

    terms related to evaluation and analysis in scientific contexts

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    78
    +0.18
    1.0%
    87
    +0.13
    0.7%
    153
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    365
    +0.18
    0.04
    482
    +0.13
    0.03
    219
    +0.11
    0.04
    Negative Logits
     outright
    -1.76
     but
    -1.73
     rather
    -1.58
    .[]{
    -1.52
     wrong
    -1.48
    ].)
    -1.45
    .";
    -1.44
    .);
    -1.43
    .
    -1.37
     if
    -1.36
    POSITIVE LOGITS
    č↵
    6.68
    ↵
    6.68
    <|outofrange|>
    6.68
    ↵
    6.68
     ↵
    6.68
    <|outofrange|>
    6.68
    ↵↵
    6.68
    ↵Č
    6.68
    <|outofrange|>
    6.68
    6.68
    Activations Density 0.239%

    No Known Activations