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. 11605
    Prev
    Next
    INDEX
    Explanations

    phrases related to support or suitability for specific purposes

    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
    358
    +0.13
    0.7%
    75
    +0.12
    0.6%
    478
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    505
    +0.13
    0.07
    358
    +0.12
    0.08
    69
    +0.10
    0.08
    Negative Logits
    grounds
    -1.54
     wrought
    -1.53
     aside
    -1.51
     flags
    -1.49
     winds
    -1.45
    er
    -1.41
     Pradesh
    -1.39
     hats
    -1.38
     indicators
    -1.37
     statutes
    -1.37
    POSITIVE LOGITS
    ice
    1.60
    thood
    1.49
     manufacture
    1.47
     use
    1.40
    li
    1.38
     Mobile
    1.36
     Use
    1.35
    deal
    1.35
     include
    1.34
     exclusively
    1.33
    Activations Density 0.599%

    No Known Activations