INDEX

Explanations

comparisons or descriptions using the word "like"

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "looks like."

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

alez

-0.90

ourse

-0.83

ESA

-0.82

Ö¼

-0.80

inion

-0.77

onding

-0.75

rief

-0.71

isexual

-0.71

ipolar

-0.70

inguished

-0.70

POSITIVE LOGITS

lier

1.08

lihood

1.05

liest

1.02

liness

0.76

 Andromeda

0.75

soDeliveryDate

0.74

flame

0.73

glers

0.70

 crap

0.66

pus

0.63

Activations Density 0.033%

comparisons or descriptions using the word "like"

instances of the phrase "looks like."

No Comments

No Known Activations