© 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. 2-RES-SM
    11. 29727
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing preferences or alternatives, particularly focusing on the word "rather."

    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
    376
    +0.11
    0.6%
    375
    +0.10
    0.6%
    74
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    375
    +0.11
    0.02
    207
    +0.10
    0.02
    165
    +0.10
    0.02
    Negative Logits
    ļ
    -2.06
    IJ
    -2.04
    Ģ
    -2.01
    Į
    -2.00
    Ļª
    -1.99
    į
    -1.91
    ľĵ
    -1.87
    ĥ
    -1.81
    º
    -1.80
    Ħ
    -1.78
    POSITIVE LOGITS
     than
    2.57
    than
    2.16
    Than
    2.03
     Than
    2.01
    bourg
    1.75
    stood
    1.68
    garten
    1.60
    keit
    1.58
    ther
    1.49
    gart
    1.47
    Activations Density 0.101%

    No Known Activations