INDEX

Explanations

nouns related to conflict or controversy

oai_token-act-pair · gpt-3.5-turbo

the prefix "un" indicating negation or opposition

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

 Pound

-0.85

Peb

-0.80

 Fields

-0.71

utical

-0.71

 tomat

-0.70

 Madden

-0.70

OPLE

-0.70

 briefs

-0.69

 Noir

-0.68

 Grayson

-0.68

POSITIVE LOGITS

fortunately

1.30

usual

1.18

iversity

1.18

likely

1.15

cles

1.11

stable

1.09

animous

1.08

expected

1.08

employment

1.07

affiliated

1.06

Activations Density 0.020%

nouns related to conflict or controversy

the prefix "un" indicating negation or opposition

No Comments

No Known Activations