INDEX

Explanations

numerical values

oai_token-act-pair · gpt-3.5-turbo

references to numerical data and catalogs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

loo

-0.90

esley

-0.68

WARD

-0.68

send

-0.67

lier

-0.66

hift

-0.65

zai

-0.65

olicy

-0.64

hips

-0.63

lyn

-0.63

POSITIVE LOGITS

ysis

0.86

antine

0.81

pty

0.81

urally

0.79

encies

0.78

ãĥĥãĥī

0.76

ogue

0.75

å¥³

0.73

urg

0.71

orically

0.71

Activations Density 0.048%

numerical values

references to numerical data and catalogs

No Comments

No Known Activations