© 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 EvalsExportsinterp-engineNEWAPI 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. 0-RES-JB
    11. 16328
    Prev
    Next
    INDEX
    Explanations

    verbs or phrases related to increasing or raising something to a higher level

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "elevated" and its variations, often in contexts related to status or levels

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    source
    -0.76
    wen
    -0.72
    ppo
    -0.72
     Guess
    -0.70
    Cos
    -0.69
    operation
    -0.69
    Mind
    -0.69
    Bee
    -0.67
    igs
    -0.67
    search
    -0.66
    POSITIVE LOGITS
     elevated
    3.70
     elev
    2.20
     elevate
    2.16
     Elev
    2.01
     elevation
    1.90
     exalted
    1.41
     heightened
    1.38
     lowered
    1.14
     enriched
    1.06
     degraded
    1.05
    Activations Density 0.021%

    No Known Activations