© 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. 27211
    Prev
    Next
    INDEX
    Explanations

    expressions of attempting or trying something

    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
    48
    +0.13
    0.7%
    489
    +0.11
    0.6%
    488
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    489
    +0.13
    0.01
    32
    +0.11
    0.02
    94
    +0.10
    0.02
    Negative Logits
    ľ
    -3.24
    ŀ
    -3.19
    «
    -3.10
    Ĺ
    -3.10
    ĥ
    -3.00
    ĺ
    -2.99
    ¼
    -2.97
    Ģ
    -2.94
    ĸ
    -2.89
    Ł
    -2.82
    POSITIVE LOGITS
     unsuccess
    1.78
    gery
    1.71
     deleting
    1.68
     nothing
    1.61
     harder
    1.61
     personally
    1.60
     google
    1.52
    ition
    1.51
     hard
    1.48
    eing
    1.47
    Activations Density 0.040%

    No Known Activations