INDEX

Explanations

phrases conveying disapproval or rebuttal

oai_token-act-pair · gpt-3.5-turbo

words related to disapproval and critique

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

 GOODMAN

-0.74

 hemor

-0.71

 Hungry

-0.68

 Insect

-0.63

 imagination

-0.63

imeters

-0.63

ivating

-0.63

 Frie

-0.62

Alz

-0.62

 intest

-0.62

POSITIVE LOGITS

ledged

0.78

stellar

0.75

essors

0.73

Dear

0.73

ials

0.71

ylon

0.67

gate

0.67

zzle

0.66

 thereof

0.66

ously

0.66

Activations Density 0.140%

phrases conveying disapproval or rebuttal

words related to disapproval and critique

No Comments

No Known Activations