INDEX

Explanations

phrases related to comparisons or contrasts

oai_token-act-pair · gpt-3.5-turbo

negative aspects or criticisms

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

 Dickinson

-0.79

 Codec

-0.71

 Richards

-0.68

 Indigo

-0.65

 Burnett

-0.65

 Powder

-0.62

 Ital

-0.59

 imprint

-0.59

 Chill

-0.59

Mod

-0.59

POSITIVE LOGITS

based

1.31

related

1.18

derived

1.15

driven

1.14

induced

1.13

style

1.13

turned

1.12

sized

1.12

only

1.11

inspired

1.10

Activations Density 0.268%

phrases related to comparisons or contrasts

negative aspects or criticisms

No Comments

No Known Activations