INDEX

Explanations

locations such as cities and countries

oai_token-act-pair · gpt-3.5-turbo

instances of the word "where," indicating a focus on locations or settings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

TAG

-0.75

âĨ

-0.73

Woman

-0.69

leeve

-0.66

hack

-0.65

termin

-0.64

apult

-0.64

fml

-0.64

ME

-0.63

Gaza

-0.63

POSITIVE LOGITS

upon

1.56

soever

0.97

 they

0.92

abouts

0.85

 temperatures

0.82

he

0.81

 residents

0.77

it

0.75

 tensions

0.74

 winters

0.74

Activations Density 0.049%

locations such as cities and countries

instances of the word "where," indicating a focus on locations or settings

No Comments

No Known Activations