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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 4-RES-JB
    11. 8509
    Prev
    Next
    INDEX
    Explanations

    phrases related to preventing negative outcomes or actions

    oai_token-act-pair · gpt-3.5-turbo

    the concept of prevention in various contexts

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

    No Comments

    Negative Logits
    ammy
    -0.84
    eah
    -0.75
    edded
    -0.69
    night
    -0.63
    emouth
    -0.61
    enegger
    -0.61
    Word
    -0.60
    ingers
    -0.60
    ATHER
    -0.59
    aturday
    -0.59
    POSITIVE LOGITS
    ative
    0.90
     inhib
    0.86
     detection
    0.85
     duplicate
    0.82
     regress
    0.76
    ministic
    0.73
     obstruct
    0.72
     allev
    0.70
    ively
    0.70
    00200000
    0.70
    Activations Density 0.027%

    No Known Activations