INDEX

Explanations

phrases related to abstract concepts or interpretations

oai_token-act-pair · gpt-3.5-turbo

phrases indicating varying interpretations or definitions of a concept

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nces

-0.81

osures

-0.74

enegger

-0.73

orst

-0.68

fam

-0.66

eches

-0.65

 Provided

-0.65

imentary

-0.62

bench

-0.62

cial

-0.62

POSITIVE LOGITS

 life

0.85

 sexuality

0.80

 morality

0.78

 masculinity

0.77

 anarchism

0.74

quickShipAvailable

0.72

 homosexuality

0.72

 Buddhism

0.72

izoph

0.71

 afterlife

0.71

Activations Density 0.271%

phrases related to abstract concepts or interpretations

phrases indicating varying interpretations or definitions of a concept

No Comments

No Known Activations