INDEX

Explanations

the word "usages" or variations of it

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'u'

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

CBI

-0.63

 removable

-0.63

 Presbyter

-0.62

 Leviathan

-0.62

 horizont

-0.62

 comma

-0.62

 resid

-0.61

 sidewalks

-0.61

 Compass

-0.60

lot

-0.59

POSITIVE LOGITS

pport

1.26

endo

1.03

ccess

1.02

cci

1.00

oti

0.94

seless

0.93

asive

0.92

asion

0.92

isine

0.91

cca

0.88

Activations Density 0.020%

the word "usages" or variations of it

occurrences of the letter 'u'

No Comments

No Known Activations