INDEX

Explanations

Well-known names or entities

oai_token-act-pair · gpt-3.5-turbo

phrases and terms associated with well-known entities or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ossession

-0.94

ascript

-0.91

alos

-0.90

plet

-0.87

©¶æ

-0.81

owder

-0.81

irlf

-0.79

otion

-0.77

illation

-0.75

rection

-0.74

POSITIVE LOGITS

 landmarks

0.74

tale

0.74

itarian

0.74

ties

0.73

stood

0.72

iary

0.69

 phenomenon

0.68

 precedent

0.68

 facts

0.67

 names

0.67

Activations Density 0.055%

Well-known names or entities

phrases and terms associated with well-known entities or concepts

No Comments

No Known Activations