INDEX

Explanations

phrases indicating a point of interest or importance in the text

oai_token-act-pair · gpt-3.5-turbo

references to unspecified or vague subjects or actions

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

âĸ

-0.67

PAC

-0.65

clus

-0.65

akedown

-0.64

 priesthood

-0.64

eworld

-0.64

egal

-0.64

NES

-0.62

ickets

-0.60

osponsors

-0.60

POSITIVE LOGITS

body

0.92

 else

0.81

Else

0.79

 Comes

0.73

 forgot

0.69

 crept

0.68

quickShipAvailable

0.68

 Werewolf

0.67

nces

0.65

witz

0.65

Activations Density 0.088%

phrases indicating a point of interest or importance in the text

references to unspecified or vague subjects or actions

No Comments

No Known Activations