INDEX

Explanations

words related to negative emotions or actions, such as "abject" or "fools."

oai_token-act-pair · gpt-3.5-turbo

words related to subjectivity and subjective language

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 cryst

-0.85

inventoryQuantity

-0.79

MpServer

-0.78

©¶æ

-0.78

 exha

-0.76

cale

-0.75

Ô

-0.73

--------------------------------------------------------

-0.72

ONSORED

-0.72

ItemThumbnailImage

-0.71

POSITIVE LOGITS

ure

1.04

ures

1.02

ions

1.00

ives

0.91

urous

0.89

ivity

0.84

utions

0.81

ution

0.81

uated

0.80

ilities

0.79

Activations Density 0.020%

words related to negative emotions or actions, such as "abject" or "fools."

words related to subjectivity and subjective language

No Comments

No Known Activations