© 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
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 2-RES-JB
    11. 3155
    Prev
    Next
    INDEX
    Explanations

    various contexts in which the word "crush" is used

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "crush" and its variations, indicating themes related to domination or overwhelm

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    gewater
    -0.86
    vernment
    -0.71
     appropri
    -0.69
    rog
    -0.68
    cale
    -0.67
    adapt
    -0.67
    gotten
    -0.65
    bis
    -0.65
     Cosponsors
    -0.64
    ported
    -0.63
    POSITIVE LOGITS
     Crush
    1.10
     crush
    1.08
    aneers
    0.70
    dragon
    0.69
    trap
    0.69
    enance
    0.68
    stals
    0.67
    avorite
    0.66
    plings
    0.65
    Boss
    0.65
    Activations Density 0.011%

    No Known Activations