INDEX

Explanations

references to prestigious achievements or positions

oai_token-act-pair · gpt-3.5-turbo

terms related to various types of awards or recognitions and events associated with them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Legends

-0.65

 TAMADRA

-0.65

 Lent

-0.65

Dot

-0.65

ynthesis

-0.64

 AMERICA

-0.64

 Blizz

-0.64

 Buffett

-0.63

 Haram

-0.62

 Hots

-0.62

POSITIVE LOGITS

ographer

1.23

keeper

1.05

wright

1.02

ologist

1.02

ographers

1.01

 dancer

0.96

icult

0.96

keepers

0.93

winner

0.92

 maker

0.90

Activations Density 0.603%

references to prestigious achievements or positions

terms related to various types of awards or recognitions and events associated with them

No Comments

No Known Activations