© 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. 7859
    Prev
    Next
    INDEX
    Explanations

    phrases related to breaking, crushing, or destruction

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "crush" and its derivatives, indicating a focus on themes of crushing or overwhelming force

    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
     appropri
    -0.72
    rog
    -0.70
    vernment
    -0.70
    nesota
    -0.69
    adapt
    -0.69
    reen
    -0.68
    iator
    -0.67
     appropriation
    -0.65
    ativity
    -0.64
    cale
    -0.63
    POSITIVE LOGITS
     crush
    0.91
     Crush
    0.90
     crushing
    0.76
    stals
    0.71
    meat
    0.71
    dragon
    0.70
    Boss
    0.69
    metal
    0.68
     earthqu
    0.68
    DOWN
    0.67
    Activations Density 0.011%

    No Known Activations