INDEX

Explanations

"full" in the context of content, such as a full list, report, story, or article

oai_token-act-pair · gpt-3.5-turbo

references to complete lists or detailed content summaries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

irgin

-0.73

Reviewer

-0.72

ucl

-0.70

 ashore

-0.68

uca

-0.67

aline

-0.65

abbit

-0.64

istors

-0.64

idated

-0.64

iques

-0.64

POSITIVE LOGITS

 text

1.11

 list

1.08

 extent

1.05

 story

0.98

 rundown

0.96

 transcript

0.96

 article

0.96

 contents

0.95

 report

0.92

text

0.92

Activations Density 0.040%

"full" in the context of content, such as a full list, report, story, or article

references to complete lists or detailed content summaries

No Comments

No Known Activations