INDEX

Explanations

occurrences of the prefix "pre-" followed by a word

oai_token-act-pair · gpt-3.5-turbo

words indicating a state of preparation or readiness

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

loo

-0.92

ãĥīãĥ©ãĤ´ãĥ³

-0.76

 GOODMAN

-0.74

ihara

-0.69

owship

-0.69

 Kingdoms

-0.68

 Darling

-0.66

 dots

-0.66

ibn

-0.65

Accessory

-0.65

POSITIVE LOGITS

cision

1.19

cedented

1.16

pared

1.13

vious

1.10

viously

1.10

hend

1.00

hens

0.98

yright

0.96

fecture

0.95

cise

0.95

Activations Density 0.007%

occurrences of the prefix "pre-" followed by a word

words indicating a state of preparation or readiness

No Comments

No Known Activations