INDEX

Explanations

phrases related to official reports or statements

oai_token-act-pair · gpt-3.5-turbo

the word "the."

oai_token-act-pair · gpt-4o-mini Triggered 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

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ò

-0.70

aba

-0.68

thood

-0.67

leeve

-0.66

 because

-0.66

ceive

-0.66

eno

-0.65

bg

-0.65

 beforehand

-0.63

POSITIVE LOGITS

 biggest

1.11

resa

1.11

oret

1.07

 oldest

1.06

odore

1.06

 simplest

1.05

 largest

1.03

 latter

1.01

 vast

1.00

 slightest

1.00

Activations Density 0.438%

phrases related to official reports or statements

the word "the."

No Comments

No Known Activations