INDEX

Explanations

negative phrases or terms used qualitatively in various contexts

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

phrases related to comparisons or contrasts

oai_token-act-pair · gpt-3.5-turbo

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

illac

-0.85

ħĭ

-0.75

¥µ

-0.73

AVG

-0.72

HCR

-0.68

 Chero

-0.67

Ń·

-0.67

 charms

-0.67

 davidjl

-0.66

 Dickinson

-0.64

POSITIVE LOGITS

advert

0.91

built

0.86

cont

0.86

 midst

0.85

depth

0.85

bred

0.83

Ear

0.82

development

0.81

state

0.80

Saharan

0.79

Activations Density 0.028%

negative phrases or terms used qualitatively in various contexts

phrases related to comparisons or contrasts

No Comments

No Known Activations