INDEX

Explanations

instances of the word "Brand" followed by a digit

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the name "Brand" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 attendant

-0.72

veyard

-0.72

pmwiki

-0.72

ROR

-0.72

fortunately

-0.69

ursed

-0.67

ancial

-0.66

SEE

-0.64

 subsistence

-0.63

ITED

-0.62

POSITIVE LOGITS

enburg

1.10

ing

0.90

Brand

0.89

 Brand

0.87

brand

0.83

ishing

0.79

eering

0.77

olph

0.76

enberg

0.74

zen

0.74

Activations Density 0.012%

instances of the word "Brand" followed by a digit

the occurrences of the name "Brand" in various contexts

No Comments

No Known Activations