INDEX

Explanations

words related to associations or connections between different concepts or entities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating association or connection between concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ynes

-0.62

NB

-0.62

anke

-0.61

Reviewed

-0.61

abases

-0.61

iger

-0.60

ucci

-0.60

 partName

-0.60

 replaces

-0.60

numbered

-0.59

POSITIVE LOGITS

 regards

0.76

stood

0.70

emonic

0.68

ryan

0.67

è¦ļéĨĴ

0.67

 extinction

0.65

ä½ľ

0.63

 Deity

0.63

 Bohem

0.60

rophic

0.60

Activations Density 0.073%

words related to associations or connections between different concepts or entities

phrases indicating association or connection between concepts

No Comments

No Known Activations