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

    Neuronpedia

    Jacobian LensNEW
    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. 3-RES-JB
    11. 5307
    Prev
    Next
    INDEX
    Explanations

    words related to avoiding or escaping something, such as evading, dodging, or escape

    oai_token-act-pair · gpt-3.5-turbo

    terms related to evasion and avoidance

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    aster
    -0.88
    onial
    -0.85
    aghan
    -0.76
    union
    -0.75
    oyal
    -0.74
    ulz
    -0.72
    strength
    -0.72
    ension
    -0.69
    RAW
    -0.68
    beer
    -0.67
    POSITIVE LOGITS
     dodge
    1.01
     evasion
    0.95
     dodging
    0.93
     evade
    0.90
     Luffy
    0.83
     dod
    0.77
    acle
    0.74
    asive
    0.72
     avoidance
    0.71
    glers
    0.70
    Activations Density 0.019%

    No Known Activations