INDEX

Explanations

phrases indicating universality or consistency across all entities

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate widespread or collective situations

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

 Siber

-0.81

uala

-0.79

 Rout

-0.67

 Berk

-0.66

Mub

-0.65

Ô

-0.64

eport

-0.63

Ads

-0.63

 Frie

-0.63

anson

-0.62

POSITIVE LOGITS

 notch

0.73

stairs

0.64

lihood

0.64

etheless

0.63

isphere

0.62

 improvement

0.62

rust

0.60

 pathological

0.59

atics

0.58

 equation

0.57

Activations Density 0.054%

phrases indicating universality or consistency across all entities

phrases that indicate widespread or collective situations

No Comments

No Known Activations