INDEX

Explanations

scientific terms and research-related phrases

oai_token-act-pair · gpt-3.5-turbo

scientific terminology related to research findings and implications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iquette

-0.75

 Franch

-0.74

 Franchise

-0.67

civ

-0.66

 Comedy

-0.66

 Drama

-0.65

 honorable

-0.63

 Titus

-0.63

 Militia

-0.62

Jimmy

-0.60

POSITIVE LOGITS

 researchers

1.08

 scientists

1.07

 biologically

0.98

 molecular

0.97

 mammalian

0.96

 biomark

0.94

 genes

0.94

 epigen

0.93

 proteins

0.92

 organisms

0.92

Activations Density 1.629%

scientific terms and research-related phrases

scientific terminology related to research findings and implications

No Comments

No Known Activations