INDEX

Explanations

The pattern "and" followed by a word or phrase

oai_token-act-pair · gpt-3.5-turbo

connectors or conjunctions used in sentences

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

assetsadobe

-0.80

 carbohyd

-0.73

quartered

-0.73

 endorsements

-0.67

avorite

-0.67

 Seym

-0.67

senal

-0.66

 Verb

-0.64

 scouting

-0.64

srf

-0.64

POSITIVE LOGITS

aman

0.89

ige

0.76

Race

0.74

allion

0.73

yg

0.72

aker

0.71

ida

0.70

ellow

0.69

UU

0.66

wed

0.66

Activations Density 0.000%

The pattern "and" followed by a word or phrase

connectors or conjunctions used in sentences

No Comments

No Known Activations