© 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. 1-RES-JB
    11. 20065
    Prev
    Next
    INDEX
    Explanations

    references to stars or celebrities

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "star"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    apons
    -1.01
    »Ĵ
    -1.00
    nsic
    -0.95
    odcast
    -0.92
    veyard
    -0.88
    berra
    -0.88
    ibaba
    -0.85
    Downloadha
    -0.85
    ĵĺ
    -0.85
    iblings
    -0.84
    POSITIVE LOGITS
     star
    1.13
     stars
    1.12
    star
    0.92
    stars
    0.92
    light
    0.85
     attraction
    0.84
    liner
    0.83
    lit
    0.78
    burst
    0.75
    lite
    0.75
    Activations Density 0.011%

    No Known Activations