INDEX

Explanations

phrases related to legal or political statements

oai_token-act-pair · gpt-3.5-turbo

ellipses or pauses in statements, indicating omitted information or a trailing thought

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

»Ĵ

-0.77

 charms

-0.74

gling

-0.70

 cradle

-0.67

grass

-0.67

åŃĲ

-0.66

 ageing

-0.66

 algae

-0.65

imore

-0.64

 Ribbon

-0.64

POSITIVE LOGITS

 âĢİ

0.90

etc

0.89

FLAG

0.84

Bloomberg

0.79

Draft

0.79

they

0.79

Freedom

0.78

emphasis

0.78

,"

0.77

 Continued

0.77

Activations Density 0.040%

phrases related to legal or political statements

ellipses or pauses in statements, indicating omitted information or a trailing thought

No Comments

No Known Activations