INDEX

Explanations

titles or headings within text

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the titles of articles, reports, or papers

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

eda

-0.79

ieri

-0.66

midt

-0.66

ometers

-0.65

aea

-0.64

EEE

-0.62

arthy

-0.62

apper

-0.61

arty

-0.61

 squared

-0.60

POSITIVE LOGITS

pins

0.82

 titled

0.79

selves

0.69

"#

0.68

icide

0.67

 misunder

0.66

("

0.65

comings

0.65

"<

0.63

nces

0.63

Activations Density 0.026%

titles or headings within text

phrases that indicate the titles of articles, reports, or papers

No Comments

No Known Activations