INDEX

Explanations

detailed descriptions of objects or scenarios

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "description"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

enthal

-0.72

 scrimmage

-0.71

acus

-0.71

nuts

-0.68

leaning

-0.68

ced

-0.66

duino

-0.65

yrus

-0.63

itsch

-0.63

inth

-0.62

POSITIVE LOGITS

 description

1.19

 descriptions

1.14

 describ

0.97

 synopsis

0.93

description

0.91

anguage

0.84

 warr

0.82

ãĤ¼ãĤ¦ãĤ¹

0.79

CRIPTION

0.78

escription

0.78

Activations Density 0.008%

detailed descriptions of objects or scenarios

repeated mentions of the word "description"

No Comments

No Known Activations