INDEX

Explanations

keywords related to news articles or written content, particularly identifying the main story

oai_token-act-pair · gpt-3.5-turbo

references to the main story or narrative in the text

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

abase

-0.82

icion

-0.75

gone

-0.73

aimon

-0.67

»Ĵ

-0.63

gem

-0.62

ignt

-0.61

ewitness

-0.59

oplan

-0.59

aiman

-0.59

POSITIVE LOGITS

Article

0.80

 ARTICLE

0.67

 reprinted

0.65

 Â»

0.62

ines

0.62

 UNCLASSIFIED

0.61

 unlocks

0.60

 Loading

0.60

Subscribe

0.60

Â¢

0.59

Activations Density 0.014%

keywords related to news articles or written content, particularly identifying the main story

references to the main story or narrative in the text

No Comments

No Known Activations