INDEX

Explanations

words related to names or personal identity

oai_token-act-pair · gpt-3.5-turbo

mentions of "IA" or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

arians

-0.68

rap

-0.67

 mustache

-0.64

alogue

-0.64

NESS

-0.63

nels

-0.63

ergy

-0.62

ary

-0.61

iple

-0.61

agher

-0.60

POSITIVE LOGITS

ppe

1.40

zza

1.28

BILITY

1.13

ppa

1.05

pered

1.04

uthor

1.03

ffe

1.03

BILITIES

1.03

zzo

1.01

ircraft

0.98

Activations Density 0.103%

words related to names or personal identity

mentions of "IA" or related terms

No Comments

No Known Activations