INDEX

Explanations

words related to exclusive or high-status items, services, or concepts

oai_token-act-pair · gpt-3.5-turbo

terms related to exclusivity and key figures associated with original ideas or plans

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

achy

-0.91

astern

-0.84

gypt

-0.77

ngth

-0.69

 Venezuel

-0.68

volent

-0.67

utic

-0.67

agues

-0.67

 Peninsula

-0.67

Myr

-0.65

POSITIVE LOGITS

ibles

0.89

ibility

0.86

ives

0.81

²¾

0.78

knit

0.77

rity

0.73

yrinth

0.73

ivity

0.73

sand

0.72

ãĥ¯ãĥ³

0.72

Activations Density 0.021%

words related to exclusive or high-status items, services, or concepts

terms related to exclusivity and key figures associated with original ideas or plans

No Comments

No Known Activations