INDEX

Explanations

words related to decrease or reduced

oai_token-act-pair · gpt-3.5-turbo

phrases related to decreases in various metrics 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

ervatives

-0.68

ansas

-0.68

dor

-0.66

wered

-0.65

Found

-0.65

Bio

-0.65

RA

-0.64

Reviewed

-0.64

ervative

-0.64

rs

-0.63

POSITIVE LOGITS

cember

0.78

 regress

0.73

 proport

0.73

 friction

0.71

imar

0.70

 inhib

0.69

ately

0.69

rences

0.68

CTR

0.68

 sidx

0.68

Activations Density 0.031%

words related to decrease or reduced

phrases related to decreases in various metrics or conditions

No Comments

No Known Activations