INDEX

Explanations

phrases indicating disapproval or opposition

oai_token-act-pair · gpt-3.5-turbo

denials and negative statements in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

azard

-0.78

utenberg

-0.75

 manually

-0.70

 bulky

-0.68

 pricey

-0.65

 Locations

-0.64

umped

-0.63

catentry

-0.63

moil

-0.63

umping

-0.62

POSITIVE LOGITS

 sane

1.07

 civilized

0.97

 acceptable

0.94

 decency

0.93

 patri

0.92

ensible

0.92

 deserve

0.92

acceptable

0.90

 remorse

0.89

 condone

0.89

Activations Density 0.552%

phrases indicating disapproval or opposition

denials and negative statements in the text

No Comments

No Known Activations