INDEX

Explanations

numeric values with a specific format

oai_token-act-pair · gpt-3.5-turbo

counts or statistics relating to social media interactions

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

okin

-0.67

wagen

-0.67

andestine

-0.66

gow

-0.66

Iterator

-0.62

arios

-0.61

ornia

-0.60

 refere

-0.59

 Gardens

-0.59

 pigeon

-0.58

POSITIVE LOGITS

/+

1.17

-+-+-+-+

1.01

/-

0.88

-+

0.83

--+

0.80

minus

0.80

=~=~

0.74

ï¸ı

0.72

 (âĪĴ

0.71

 Spoiler

0.71

Activations Density 0.106%

numeric values with a specific format

counts or statistics relating to social media interactions

No Comments

No Known Activations