© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 11696
    Prev
    Next
    INDEX
    Explanations

    references to video content and related media

    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
    292
    +0.13
    0.7%
    363
    +0.13
    0.7%
    307
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    363
    +0.13
    0.03
    354
    +0.13
    0.03
    151
    +0.11
    0.02
    Negative Logits
    hip
    -2.51
    ľĵ
    -2.16
    hips
    -1.91
    dern
    -1.88
    udson
    -1.81
    ¯
    -1.76
    chaft
    -1.64
    dere
    -1.61
    shire
    -1.59
    pshire
    -1.56
    POSITIVE LOGITS
     clips
    3.05
     footage
    2.80
     games
    2.45
     clip
    2.40
     playback
    2.31
     tutorials
    2.31
     streaming
    2.13
     codec
    2.10
     tapes
    2.07
     game
    2.04
    Activations Density 0.227%

    No Known Activations