INDEX

Explanations

mentions of serious or critical situations, often related to danger or risk

oai_token-act-pair · gpt-3.5-turbo

instances of the word "grave" and its variations, indicating serious or severe situations

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

£ı

-0.71

ivity

-0.69

ILA

-0.68

icer

-0.67

ĪĴ

-0.65

 GOODMAN

-0.64

ĩ

-0.64

Hyper

-0.64

æ©Ł

-0.62

ynthesis

-0.62

POSITIVE LOGITS

lly

1.42

grave

1.09

fur

0.93

ly

0.90

 grave

0.90

faced

0.87

stones

0.87

 robbing

0.85

bed

0.84

stone

0.81

Activations Density 0.014%

mentions of serious or critical situations, often related to danger or risk

instances of the word "grave" and its variations, indicating serious or severe situations

No Comments

No Known Activations