INDEX

Explanations

particular phrases to prompt the reader to continue reading

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reading" and its variations

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

»Ĵ

-0.80

ĪĴ

-0.79

ignt

-0.73

hap

-0.73

opard

-0.72

footed

-0.70

pora

-0.69

given

-0.68

 Gleaming

-0.67

esville

-0.66

POSITIVE LOGITS

...]

0.80

 toggle

0.75

 WATCHED

0.70

til

0.65

 âĨĴ

0.64

 pauses

0.63

 Expand

0.62

âĢ¦]

0.61

 unsupported

0.61

 presses

0.61

Activations Density 0.011%

particular phrases to prompt the reader to continue reading

instances of the word "reading" and its variations

No Comments

No Known Activations