© 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. 0-RES-JB
    11. 1138
    Prev
    Next
    INDEX
    Explanations

    phrases related to belief or conviction

    oai_token-act-pair · gpt-3.5-turbo

    expressions related to belief and trust

    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
    rawl
    -0.65
     Interface
    -0.60
    iol
    -0.60
     cleanup
    -0.59
    cture
    -0.59
     overhead
    -0.58
    arc
    -0.57
     interface
    -0.56
     breakdown
    -0.56
     ++
    -0.56
    POSITIVE LOGITS
     believing
    3.53
     trusting
    1.83
     fearing
    1.71
     belief
    1.65
     believe
    1.45
     knowing
    1.45
     disbel
    1.44
     believer
    1.41
     expecting
    1.40
     believed
    1.35
    Activations Density 0.011%

    No Known Activations