INDEX

Explanations

names of organizations, courts, or legal entities

oai_token-act-pair · gpt-3.5-turbo

prominent capitalized terms and acronyms related to organizations or entities

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

 teasp

-0.76

theless

-0.76

Ô

-0.68

 acknow

-0.66

 sqor

-0.63

 tiss

-0.63

 pione

-0.63

 eleph

-0.61

é¾įå¥ĳå£«

-0.61

ñ

-0.60

POSITIVE LOGITS

ibrarian

0.75

agame

0.72

mbudsman

0.72

osphere

0.71

ariat

0.70

vre

0.70

yssey

0.70

isphere

0.70

ospels

0.69

problem

0.67

Activations Density 0.185%

names of organizations, courts, or legal entities

prominent capitalized terms and acronyms related to organizations or entities

No Comments

No Known Activations