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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 28724
    Prev
    Next
    INDEX
    Explanations

    references to hate or references to specific locations or events connected to it

    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
    156
    +0.48
    3.2%
    111
    +0.36
    2.4%
    294
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.48
    0.28
    111
    +0.36
    0.29
    71
    +0.10
    0.25
    Negative Logits
    cern
    -1.30
     WHETHER
    -1.29
    liers
    -1.28
     Else
    -1.26
     2020
    -1.26
     transparent
    -1.23
    except
    -1.19
     forthcoming
    -1.18
    iom
    -1.16
     2019
    -1.14
    POSITIVE LOGITS
    .’”
    1.47
    iston
    1.46
     pylori
    1.44
     insult
    1.41
    umor
    1.38
    emp
    1.31
     acres
    1.28
    ĻĤ
    1.27
    acker
    1.26
    bard
    1.26
    Activations Density 0.329%

    No Known Activations