INDEX

Explanations

phrases indicating that something is no longer the case or valid

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the cessation or termination of states, roles, or conditions

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

 nonetheless

-0.87

 sqor

-0.74

 nevertheless

-0.74

eport

-0.73

 shrew

-0.68

yip

-0.66

mac

-0.66

anan

-0.65

igl

-0.62

iscal

-0.61

POSITIVE LOGITS

 anymore

1.04

nor

0.89

Reloaded

0.74

thumbnails

0.67

orie

0.67

 Canaver

0.67

ORED

0.66

ONSORED

0.65

 idle

0.65

steen

0.64

Activations Density 0.113%

phrases indicating that something is no longer the case or valid

phrases indicating the cessation or termination of states, roles, or conditions

No Comments

No Known Activations