© 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. 27423
    Prev
    Next
    INDEX
    Explanations

    references to authors and titles in a bibliographic context

    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
    495
    +0.12
    0.7%
    183
    +0.12
    0.7%
    184
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    183
    +0.12
    0.02
    184
    +0.12
    0.02
    410
    +0.11
    0.03
    Negative Logits
    equals
    -1.59
    )",
    -1.46
    ?)
    -1.44
    psy
    -1.39
    ?",
    -1.39
    tat
    -1.35
    gly
    -1.34
    formatter
    -1.34
    inese
    -1.33
    alloc
    -1.32
    POSITIVE LOGITS
    Į
    2.82
    ĻĤ
    2.65
    ¿½
    2.28
    ľĵ
    2.18
    ³
    1.92
    IJ
    1.84
    ¬
    1.83
    °
    1.83
    ´
    1.82
    İ
    1.80
    Activations Density 0.135%

    No Known Activations