INDEX

Explanations

phrases related to taking responsibility and expressing beliefs

oai_token-act-pair · gpt-3.5-turbo

expressions of personal knowledge and understanding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

éĹĺ

-0.66

ournal

-0.63

ccording

-0.63

 subur

-0.61

 Stores

-0.60

 Banana

-0.59

 Santiago

-0.59

atown

-0.58

Caf

-0.56

yip

-0.56

POSITIVE LOGITS

 accordingly

1.27

 therein

0.94

 thereof

0.94

 thereafter

0.92

 consequently

0.90

 thereto

0.89

 therefore

0.82

 thankful

0.80

 glad

0.80

 afterwards

0.76

Activations Density 0.627%

phrases related to taking responsibility and expressing beliefs

expressions of personal knowledge and understanding

No Comments

No Known Activations