INDEX

Explanations

images or photographs mentioned in text

oai_token-act-pair · gpt-3.5-turbo

instances of the word "picture" or its variants

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

inventoryQuantity

-0.74

claimer

-0.73

Ì¶

-0.65

$$

-0.65

conduct

-0.65

margin

-0.62

hyde

-0.62

 stem

-0.62

 effic

-0.61

die

-0.61

POSITIVE LOGITS

 Pict

1.29

ographs

1.23

Pict

1.14

ured

1.03

ures

1.02

ograms

1.01

orial

1.00

urers

0.98

ogram

0.93

anooga

0.90

Activations Density 0.004%

images or photographs mentioned in text

instances of the word "picture" or its variants

No Comments

No Known Activations