INDEX

Explanations

phrases related to product launches and announcements

oai_token-act-pair · gpt-3.5-turbo

phrasal structures indicating the use or classification of items or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Â·Â·

-0.67

ivas

-0.66

aghan

-0.65

iger

-0.65

ersen

-0.61

MENTS

-0.61

ãĤ¤ãĥĪ

-0.61

lah

-0.61

rol

-0.61

izabeth

-0.60

POSITIVE LOGITS

ynchron

1.16

pires

1.14

 part

1.13

pired

1.04

part

0.95

 standalone

0.89

an

0.86

 follows

0.82

0.80

semble

0.79

Activations Density 0.110%

phrases related to product launches and announcements

phrasal structures indicating the use or classification of items or concepts

No Comments

No Known Activations