INDEX

Explanations

credit-related terms or mentions

oai_token-act-pair · gpt-3.5-turbo

references to credit cards and related financial terminology

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

Bei

-0.78

Osw

-0.76

vae

-0.74

gha

-0.67

xy

-0.66

ãĥĥãĥĪ

-0.66

++++++++

-0.65

ãĤ«

-0.65

 Lans

-0.64

Wem

-0.64

POSITIVE LOGITS

 card

1.10

card

1.07

worthiness

1.02

 CARD

1.00

cards

0.94

worthy

0.92

enza

0.92

 cards

0.91

 Cards

0.87

Card

0.85

Activations Density 0.015%

credit-related terms or mentions

references to credit cards and related financial terminology

No Comments

No Known Activations