INDEX

Explanations

references to the term "King"

oai_token-act-pair · gpt-3.5-turbo

references to the word "King."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

eria

-0.81

TING

-0.78

ters

-0.77

uated

-0.73

ATIONAL

-0.73

ciplinary

-0.72

ename

-0.71

Ö¼

-0.70

ission

-0.69

arial

-0.68

POSITIVE LOGITS

pin

1.26

uin

1.15

doms

1.13

pins

1.11

dom

1.10

fish

1.01

DOM

0.97

 Kong

0.94

stown

0.90

 Clancy

0.89

Activations Density 0.014%

references to the term "King"

references to the word "King."

No Comments

No Known Activations