INDEX

Explanations

words with the prefix "en-" followed by a verb or noun

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "en" in various forms

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

DragonMagazine

-0.98

ĳå£«

-0.71

 Palmer

-0.70

hyde

-0.69

ramid

-0.68

TPPStreamerBot

-0.66

 Murd

-0.64

Ĥª

-0.64

 Orbit

-0.62

ocene

-0.62

POSITIVE LOGITS

viron

1.34

joined

1.08

unci

1.08

semb

1.03

chanted

0.99

 masse

0.99

cyclop

0.98

igma

0.98

forcer

0.98

emy

0.96

Activations Density 0.012%

words with the prefix "en-" followed by a verb or noun

instances of the prefix "en" in various forms

No Comments

No Known Activations