INDEX

Explanations

phrases starting with "Re" followed by a single word

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term 'Re', suggesting a focus on references or reactions in the text

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Pebble

-0.72

åŃ

-0.72

 buckle

-0.68

 à¨

-0.68

DragonMagazine

-0.67

ãĥ¼ãĥĨãĤ£

-0.66

£ı

-0.65

 GEAR

-0.65

ãģĨ

-0.65

ãĥ¼ãĥĨ

-0.65

POSITIVE LOGITS

becca

1.27

ceived

1.18

levant

1.17

verend

1.07

venge

1.07

asonable

1.05

fect

1.03

ceiver

1.02

ggie

1.02

jected

1.02

Activations Density 0.024%

phrases starting with "Re" followed by a single word

occurrences of the term 'Re', suggesting a focus on references or reactions in the text

No Comments

No Known Activations