INDEX

Explanations

names with professions, titles, or organizations

oai_token-act-pair · gpt-3.5-turbo

URLs or web-related references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wagen

-0.98

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.77

Ĥİ

-0.74

illac

-0.71

yang

-0.69

Versions

-0.69

paralle

-0.68

Pie

-0.67

matically

-0.65

yz

-0.65

POSITIVE LOGITS

 Frontier

0.75

AFP

0.69

mare

0.67

 Shutterstock

0.67

 COURT

0.66

 reader

0.65

ms

0.64

www

0.64

anwhile

0.63

 pseudonym

0.63

Activations Density 0.052%

names with professions, titles, or organizations

URLs or web-related references

No Comments

No Known Activations