INDEX

Explanations

articles ('a', 'an', 'the') followed by a noun

oai_token-act-pair · gpt-3.5-turbo

articles and determiners like "a" and "an"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Lodge

-0.78

Veg

-0.77

antle

-0.76

(#

-0.75

kamp

-0.72

 Niet

-0.71

 Kris

-0.68

mys

-0.68

chuk

-0.67

 Mens

-0.66

POSITIVE LOGITS

usterity

1.28

etheless

1.13

theless

0.93

onymous

0.82

 sudden

0.79

bsite

0.78

uckland

0.78

cknow

0.78

manent

0.77

mosp

0.73

Activations Density 0.144%

articles ('a', 'an', 'the') followed by a noun

articles and determiners like "a" and "an"

No Comments

No Known Activations