INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to U.S. states and cities

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

istration

-0.80

lihood

-0.79

direction

-0.76

ullivan

-0.76

ashington

-0.75

phabet

-0.75

inance

-0.75

displayText

-0.74

acement

-0.71

Support

-0.71

POSITIVE LOGITS

 Nights

1.08

 winters

1.03

 natives

1.01

 Avenue

0.96

 Skies

0.96

 skyline

0.95

 accents

0.94

 accent

0.93

 Horror

0.85

Ave

0.85

Activations Density 0.437%

locations or places

references to U.S. states and cities

No Comments

No Known Activations