INDEX

Explanations

single-word occurrences related to fundamental or essential concepts

oai_token-act-pair · gpt-3.5-turbo

references to fundamental or essential concepts, principles, or needs

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Avoid

-0.78

Chance

-0.77

/-

-0.73

âĹ¼

-0.69

 Avoid

-0.68

 Boards

-0.65

ï¸

-0.65

kers

-0.65

 Mouth

-0.64

Drug

-0.64

POSITIVE LOGITS

basic

1.18

 necessities

1.08

 basic

1.02

 arithmetic

0.94

 basics

0.93

 tenets

0.86

Basic

0.86

 essentials

0.85

 components

0.81

essential

0.80

Activations Density 0.009%

single-word occurrences related to fundamental or essential concepts

references to fundamental or essential concepts, principles, or needs

No Comments

No Known Activations