INDEX

Explanations

phrases containing the word "note"

oai_token-act-pair · gpt-3.5-turbo

the repeated mention of the term "note."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

£ı

-0.77

Ļ½

-0.74

dimension

-0.69

 seat

-0.67

 tender

-0.67

 mathemat

-0.64

 reconc

-0.64

 dred

-0.63

 retention

-0.61

 overrun

-0.61

POSITIVE LOGITS

ote

1.25

OPLE

1.20

otes

1.15

tsky

0.95

ito

0.91

ople

0.89

lete

0.89

chnology

0.89

chn

0.89

ague

0.85

Activations Density 0.011%

phrases containing the word "note"

the repeated mention of the term "note."

No Comments

No Known Activations