INDEX

Explanations

phrases related to physical suffering and hardship

oai_token-act-pair · gpt-3.5-turbo

instances of words ending with 'ings'

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

 earthqu

-0.94

ISON

-0.70

SIGN

-0.68

isons

-0.65

utherford

-0.63

PDATE

-0.63

ILD

-0.62

ERA

-0.62

VER

-0.60

vier

-0.59

POSITIVE LOGITS

hots

1.23

omething

1.08

peed

1.07

pace

1.07

cape

1.05

tons

1.03

hot

1.00

creen

0.98

uits

0.94

lash

0.92

Activations Density 0.039%

phrases related to physical suffering and hardship

instances of words ending with 'ings'

No Comments

No Known Activations