Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29586
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the provision of information or services

    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
    217
    +0.13
    0.7%
    246
    +0.11
    0.6%
    185
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    217
    +0.13
    0.04
    303
    +0.11
    0.04
    500
    +0.10
    0.04
    Negative Logits
    isme
    -1.65
    Bigr
    -1.48
    enson
    -1.45
    hors
    -1.44
    TOR
    -1.41
    LINE
    -1.40
    ullah
    -1.37
    hill
    -1.36
    ÑĤа
    -1.36
    unched
    -1.35
    POSITIVE LOGITS
     outstanding
    1.85
    eful
    1.58
    iture
    1.56
     adequate
    1.51
     invaluable
    1.47
     excellent
    1.44
    asive
    1.42
     supplementary
    1.42
     sender
    1.38
     satisfactory
    1.38
    Activations Density 0.126%

    No Known Activations