INDEX

Explanations

references or quotations from sources

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate confirmation or attribution of information

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

FACE

-0.85

enegger

-0.83

TAG

-0.82

Interstitial

-0.78

wagen

-0.77

oshenko

-0.75

ACP

-0.74

Beat

-0.70

schild

-0.69

Shell

-0.68

POSITIVE LOGITS

 Medline

0.78

UA

0.72

ISI

0.70

 wealth

0.67

ortium

0.66

 cite

0.66

 Mund

0.65

FF

0.65

 Levant

0.64

ucl

0.63

Activations Density 0.284%

references or quotations from sources

phrases that indicate confirmation or attribution of information

No Comments

No Known Activations