INDEX

Explanations

descriptive language related to imagery

oai_token-act-pair · gpt-3.5-turbo

concepts related to visual imagery and symbolism in various contexts

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

clud

-0.83

lad

-0.73

Tax

-0.70

galitarian

-0.64

ership

-0.64

Journal

-0.63

bern

-0.63

tein

-0.63

erm

-0.62

gres

-0.60

POSITIVE LOGITS

 imagery

0.92

uggest

0.83

icz

0.81

ickr

0.78

abwe

0.75

icol

0.72

ays

0.71

 destro

0.70

seys

0.70

anes

0.70

Activations Density 0.055%

descriptive language related to imagery

concepts related to visual imagery and symbolism in various contexts

No Comments

No Known Activations