INDEX

Explanations

proper nouns related to individuals or entities, likely named "Norris" and "iba"

oai_token-act-pair · gpt-3.5-turbo

names of notable individuals, particularly those related to awards or criminal cases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ebook

-0.84

izations

-0.72

plane

-0.70

uers

-0.70

izoph

-0.69

pt

-0.68

tails

-0.67

iste

-0.66

planes

-0.66

framework

-0.65

POSITIVE LOGITS

 Norris

1.04

Downloadha

0.81

daq

0.80

 Brothers

0.78

cliffe

0.76

 Jarvis

0.73

BuyableInstoreAndOnline

0.71

 Bros

0.70

wick

0.69

 mund

0.68

Activations Density 0.011%

proper nouns related to individuals or entities, likely named "Norris" and "iba"

names of notable individuals, particularly those related to awards or criminal cases

No Comments

No Known Activations