INDEX

Explanations

names or terms related to winners or successful outcomes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "win" and related terms connected to it

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

actionDate

-0.87

arij

-0.72

channelAvailability

-0.69

Īè

-0.65

anni

-0.65

 pockets

-0.64

è¦ļéĨĴ

-0.64

ãĥķãĤ©

-0.61

 deviations

-0.60

 reneg

-0.60

POSITIVE LOGITS

stein

0.83

wagen

0.82

bilt

0.77

 Ferry

0.75

cha

0.73

çİĭ

0.73

ipeg

0.71

lich

0.70

wald

0.69

Cola

0.68

Activations Density 0.206%

names or terms related to winners or successful outcomes

instances of the word "win" and related terms connected to it

No Comments

No Known Activations