INDEX

Explanations

adjectives related to following rules or regulations

oai_token-act-pair · gpt-3.5-turbo

references to law-abiding individuals and their actions

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

hy

-0.77

oult

-0.73

 Hosp

-0.70

ammy

-0.70

kill

-0.68

ilk

-0.67

cair

-0.66

Nou

-0.65

olog

-0.65

tips

-0.63

POSITIVE LOGITS

abiding

1.41

 abiding

1.08

ĸļ

0.86

 compr

0.75

obe

0.75

withstanding

0.73

ctuary

0.73

¿½

0.73

rats

0.71

ignt

0.70

Activations Density 0.011%

adjectives related to following rules or regulations

references to law-abiding individuals and their actions

No Comments

No Known Activations