INDEX

Explanations

numerical values associated with monetary figures

oai_token-act-pair · gpt-3.5-turbo

phrases related to financial metrics expressed in a "per" format

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Parties

-0.73

PT

-0.66

="#

-0.63

 Bowie

-0.62

 Tests

-0.60

 Password

-0.60

hess

-0.59

tl

-0.59

 Phar

-0.59

Eug

-0.58

POSITIVE LOGITS

 year

1.02

ann

0.98

 capita

0.89

 month

0.87

decade

0.80

year

0.80

month

0.78

 annually

0.76

yr

0.76

 annual

0.76

Activations Density 0.068%

numerical values associated with monetary figures

phrases related to financial metrics expressed in a "per" format

No Comments

No Known Activations