INDEX

Explanations

words related to geographical locations or entities

oai_token-act-pair · gpt-3.5-turbo

geographical names and locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 warranted

-0.69

 censored

-0.65

 complete

-0.63

 Fallout

-0.63

 insecure

-0.62

 perfect

-0.62

Âł

-0.62

ADD

-0.61

 logically

-0.61

 depressed

-0.61

POSITIVE LOGITS

ovo

1.19

inka

1.17

pora

1.16

je

1.15

hus

1.15

abad

1.12

vik

1.12

wana

1.12

unda

1.10

uku

1.08

Activations Density 0.403%

words related to geographical locations or entities

geographical names and locations

No Comments

No Known Activations