INDEX

Explanations

words related to moral values and principles

oai_token-act-pair · gpt-3.5-turbo

plural nouns or adjectives and words related to groups or categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

PET

-0.79

amba

-0.70

Tex

-0.66

INO

-0.66

UL

-0.63

UCK

-0.62

window

-0.61

LECT

-0.60

ATED

-0.59

Sharp

-0.59

POSITIVE LOGITS

gemony

0.90

ashtra

0.80

rahim

0.78

apego

0.78

ndra

0.73

brids

0.73

ths

0.72

ilege

0.71

zsche

0.70

andum

0.70

Activations Density 0.221%

words related to moral values and principles

plural nouns or adjectives and words related to groups or categories

No Comments

No Known Activations