INDEX

Explanations

phrases related to impersonation and fraud

oai_token-act-pair · gpt-3.5-turbo

references to impersonation and related terms

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

vier

-0.82

OPLE

-0.74

oms

-0.74

 Wester

-0.73

 RELEASE

-0.64

 Homs

-0.64

 Relief

-0.63

 Clement

-0.63

Hen

-0.62

velt

-0.61

POSITIVE LOGITS

 imperson

0.94

apixel

0.77

rity

0.73

lar

0.72

daq

0.72

raltar

0.72

orpor

0.72

metics

0.71

raphics

0.71

 spoof

0.70

Activations Density 0.039%

phrases related to impersonation and fraud

references to impersonation and related terms

No Comments

No Known Activations