INDEX

Explanations

references to organizations or groups with the acronym "AN."

oai_token-act-pair · gpt-3.5-turbo

the term "AN" or references to actions associated with it

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĤ¡

-0.74

 Cind

-0.73

lda

-0.72

 Wink

-0.70

ï¸ı

-0.68

lahoma

-0.67

 guiIcon

-0.66

odore

-0.66

mented

-0.66

Kem

-0.65

POSITIVE LOGITS

igans

1.16

alyst

1.01

ARCH

0.99

OVA

0.98

NING

0.95

ufact

0.94

NER

0.90

OTHER

0.88

ESE

0.86

IST

0.86

Activations Density 0.013%

references to organizations or groups with the acronym "AN."

the term "AN" or references to actions associated with it

No Comments

No Known Activations