INDEX

Explanations

words related to heightened states or intensification

oai_token-act-pair · gpt-3.5-turbo

language related to heightened or intensified conditions and surveillance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

å£«

-0.75

µ

-0.74

eh

-0.71

iop

-0.70

OH

-0.69

linux

-0.69

NCT

-0.68

ships

-0.65

é¾įå

-0.65

į

-0.65

POSITIVE LOGITS

 heightened

1.16

 unbeliev

1.06

 enthusi

0.95

 sensitivity

0.94

 referen

0.93

 awareness

0.92

 guiActiveUn

0.90

 tremend

0.89

 scrutiny

0.88

 alert

0.87

Activations Density 0.012%

words related to heightened states or intensification

language related to heightened or intensified conditions and surveillance

No Comments

No Known Activations