INDEX

Explanations

instances of the word "see" or "saw" followed by an object

oai_token-act-pair · gpt-3.5-turbo

phrases related to perception or observation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lator

-0.73

Newsletter

-0.67

 Researchers

-0.63

inion

-0.63

PHOTOS

-0.63

quer

-0.62

 ILCS

-0.62

\<

-0.62

ulum

-0.61

EStreamFrame

-0.60

POSITIVE LOGITS

fit

1.37

 parallels

1.29

 firsthand

1.05

 similarities

1.04

eye

0.97

 opportunity

0.96

 visions

0.95

 possibilities

0.90

 opportunities

0.89

 merit

0.87

Activations Density 0.103%

instances of the word "see" or "saw" followed by an object

phrases related to perception or observation

No Comments

No Known Activations