INDEX

Explanations

phrases or acronyms related to organizations or events

oai_token-act-pair · gpt-3.5-turbo

references to financial information or funding entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

assian

-0.79

many

-0.72

agically

-0.70

 Siren

-0.69

 gorilla

-0.68

taker

-0.67

ographer

-0.67

nia

-0.67

ocrats

-0.66

azines

-0.66

POSITIVE LOGITS

ATURES

1.05

FE

1.03

VE

1.02

ruary

1.01

ATURE

0.99

FE

0.98

ET

0.98

EMBER

0.89

VER

0.89

BR

0.87

Activations Density 0.008%

phrases or acronyms related to organizations or events

references to financial information or funding entities

No Comments

No Known Activations