INDEX

Explanations

noun phrases related to opening or beginnings

oai_token-act-pair · gpt-3.5-turbo

references to opening or revealing something significant

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

eger

-0.70

jad

-0.70

omorphic

-0.64

kov

-0.64

pell

-0.61

chem

-0.60

Kings

-0.60

axe

-0.59

kh

-0.59

FactoryReloaded

-0.59

POSITIVE LOGITS

 doors

0.98

 Pandora

0.94

 door

0.83

 Doors

0.83

 drawer

0.82

 backdoor

0.77

izons

0.75

 valves

0.74

 portal

0.74

uberty

0.72

Activations Density 0.099%

noun phrases related to opening or beginnings

references to opening or revealing something significant

No Comments

No Known Activations