INDEX

Explanations

words related to the term "Sw"

oai_token-act-pair · gpt-3.5-turbo

references to specific entities or names associated with the term "Sw."

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

OPLE

-0.76

============

-0.70

 theater

-0.70

EStream

-0.68

 preferential

-0.66

âĸ¬âĸ¬

-0.66

66666666

-0.65

ä»£

-0.65

BILITY

-0.65

 envy

-0.64

POSITIVE LOGITS

addle

1.19

imming

1.16

immer

1.16

ifty

1.14

indle

1.14

ollen

1.12

itched

1.12

allow

1.11

inging

1.11

artz

1.11

Activations Density 0.011%

words related to the term "Sw"

references to specific entities or names associated with the term "Sw."

No Comments

No Known Activations