INDEX

Explanations

proper nouns within specific contexts related to news articles or reports

oai_token-act-pair · gpt-3.5-turbo

phrases containing the word "of"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

accompan

-0.82

 includ

-0.72

userc

-0.70

isol

-0.70

respective

-0.70

 relate

-0.67

TEXT

-0.63

 amount

-0.62

netflix

-0.61

Gallery

-0.61

POSITIVE LOGITS

 Syracuse

0.95

 Georgetown

0.92

 Bellev

0.90

 Honolulu

0.85

 Princeton

0.85

 Harvard

0.85

 Philadelphia

0.85

 Jacksonville

0.84

 Providence

0.83

 Auburn

0.83

Activations Density 0.073%

proper nouns within specific contexts related to news articles or reports

phrases containing the word "of"

No Comments

No Known Activations