INDEX

Explanations

phrases related to legal and investigative actions, cautionary statements, and negative assertions about honesty and integrity

oai_token-act-pair · gpt-3.5-turbo

warnings or cautions related to legal or social issues

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

ggles

-0.74

ortium

-0.72

ishable

-0.67

 Pinball

-0.66

ibaba

-0.66

inav

-0.66

escent

-0.63

eworld

-0.63

undrum

-0.62

erenn

-0.62

POSITIVE LOGITS

 unfairly

0.75

 outwe

0.74

 ineffective

0.73

 misinterpret

0.72

instead

0.67

entimes

0.67

Coun

0.66

 instead

0.66

 retribution

0.65

Instead

0.64

Activations Density 1.771%

phrases related to legal and investigative actions, cautionary statements, and negative assertions about honesty and integrity

warnings or cautions related to legal or social issues

No Comments

No Known Activations