© 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. 4-RES-SM
    11. 13384
    Prev
    Next
    INDEX
    Explanations

    structured data or metadata elements within documents

    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
    307
    +0.18
    1.2%
    78
    +0.18
    1.1%
    466
    +0.16
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    78
    +0.18
    0.01
    9
    +0.18
    0.06
    119
    +0.16
    0.06
    Negative Logits
    itrile
    -1.74
    unnumbered
    -1.69
    plete
    -1.44
    ize
    -1.28
     modo
    -1.24
    ify
    -1.23
     times
    -1.22
    ocarcinoma
    -1.21
     onc
    -1.21
    ubes
    -1.20
    POSITIVE LOGITS
    Ĵ
    2.97
    Īĺ
    2.94
    ĵ
    2.89
    ŀ
    2.82
    2.81
    ↵↵
    2.81
    č↵
    2.81
    <|outofrange|>
    2.81
    2.81
    2.81
    Activations Density 0.455%

    No Known Activations