INDEX

Explanations

miniature versions of objects or concepts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "mini" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

âķĲâķĲ

-0.74

Ö¼

-0.70

lain

-0.69

yer

-0.66

lines

-0.63

rite

-0.62

 Grind

-0.62

addock

-0.62

xual

-0.62

iquette

-0.61

POSITIVE LOGITS

ature

1.11

atur

1.05

ocular

0.91

atures

0.87

 mini

0.76

oses

0.74

istered

0.73

 sized

0.73

ota

0.72

oby

0.72

Activations Density 0.006%

miniature versions of objects or concepts

occurrences of the word "mini" in various contexts

No Comments

No Known Activations