© 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. 2-RES-JB
    11. 9392
    Prev
    Next
    INDEX
    Explanations

    words related to physical fractures or weaknesses

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "crack" in various contexts

    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
    ikk
    -0.82
     Leth
    -0.78
    urse
    -0.65
    ucky
    -0.64
     Pwr
    -0.63
     Jinn
    -0.62
     Fenrir
    -0.61
    yss
    -0.60
    phis
    -0.58
     Rath
    -0.58
    POSITIVE LOGITS
    pots
    1.15
    ible
    0.99
    Berry
    0.92
    ibly
    0.90
    acies
    0.85
    lings
    0.84
    pot
    0.83
    buster
    0.80
     cracking
    0.79
    bowl
    0.79
    Activations Density 0.006%

    No Known Activations