© 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. 30947
    Prev
    Next
    INDEX
    Explanations

    phrases indicating user instructions or commands

    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
    95
    +0.19
    1.0%
    342
    +0.17
    1.0%
    459
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.19
    0.02
    459
    +0.17
    0.04
    95
    +0.14
    0.03
    Negative Logits
    ĻĤ
    -3.34
    ľ
    -2.92
    Ŀ
    -2.91
    ĭ
    -2.83
    ŀ
    -2.64
    µ
    -2.60
    °
    -2.58
    ĺ
    -2.57
    «
    -2.55
    Ĭ
    -2.54
    POSITIVE LOGITS
    indexOf
    1.68
    **:
    1.62
    :`
    1.59
     wont
    1.58
     vs
    1.55
     Answer
    1.54
    haven
    1.53
    getString
    1.51
    stringify
    1.48
    fiddle
    1.47
    Activations Density 0.358%

    No Known Activations