INDEX

Explanations

words related to employment and working

oai_token-act-pair · gpt-3.5-turbo

references to employment or labor-related activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ylon

-0.72

anamo

-0.72

 Flavoring

-0.69

wcs

-0.68

 Cricket

-0.65

EStream

-0.65

asar

-0.64

————

-0.62

 Bubble

-0.61

 Predators

-0.60

POSITIVE LOGITS

bench

1.16

 ethic

1.15

station

1.06

flows

1.05

manship

1.04

hops

0.98

work

0.91

mate

0.85

horse

0.84

fare

0.78

Activations Density 0.069%

words related to employment and working

references to employment or labor-related activities

No Comments

No Known Activations