INDEX

Explanations

the word "neither" followed by a mention of two contrasting elements

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the absence or negation of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Vers

-0.77

eds

-0.75

psc

-0.74

ped

-0.69

inav

-0.67

ortium

-0.66

bows

-0.65

edu

-0.65

soType

-0.65

olid

-0.64

POSITIVE LOGITS

icably

0.75

theless

0.74

 percept

0.71

 overtly

0.70

ĸļ

0.68

soever

0.67

ĨĴ

0.66

[_

0.65

 sexes

0.65

osaurus

0.63

Activations Density 0.006%

the word "neither" followed by a mention of two contrasting elements

phrases indicating the absence or negation of something

No Comments

No Known Activations