INDEX

Explanations

words with the suffix '-y' followed by a strong activation value, particularly 'y' itself

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'y'

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

raltar

-0.81

 Examiner

-0.78

IBLE

-0.70

icably

-0.68

insula

-0.68

itures

-0.67

bernatorial

-0.65

ãĥ¯

-0.63

vou

-0.63

 foss

-0.63

POSITIVE LOGITS

ield

1.02

ielding

0.99

Å«

0.92

aku

0.90

ank

0.86

olk

0.85

mbol

0.81

ikes

0.81

ng

0.80

STEM

0.80

Activations Density 0.055%

words with the suffix '-y' followed by a strong activation value, particularly 'y' itself

occurrences of the letter 'y'

No Comments

No Known Activations