INDEX

Explanations

phrases indicating collective responsibility or unity

oai_token-act-pair · gpt-3.5-turbo

references to collective experiences or shared knowledge

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

•All the Features!

No Comments

Negative Logits

luster

-0.75

ahime

-0.61

 arresting

-0.61

ESH

-0.60

ALL

-0.60

folios

-0.58

erate

-0.58

mination

-0.57

istant

-0.57

lash

-0.56

POSITIVE LOGITS

ocating

1.09

 alike

0.90

 agree

0.87

 kinds

0.85

 collectively

0.83

uding

0.78

################

0.77

 agreed

0.77

 gotta

0.72

together

0.72

Activations Density 0.034%

phrases indicating collective responsibility or unity

references to collective experiences or shared knowledge

No Comments

No Known Activations