INDEX

Explanations

references to specific names, particularly "Doyle", "Sauron", and "Capcom"

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to notable figures and franchises

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

atile

-0.92

atility

-0.87

rontal

-0.86

ighting

-0.85

uration

-0.82

urities

-0.71

ernand

-0.71

ificent

-0.71

ilater

-0.70

urations

-0.70

POSITIVE LOGITS

 Doyle

1.02

liest

0.80

NRS

0.74

IVERS

0.72

IRO

0.72

Wr

0.72

Wheel

0.71

hiba

0.68

evil

0.67

wolves

0.65

Activations Density 0.023%

references to specific names, particularly "Doyle", "Sauron", and "Capcom"

proper nouns related to notable figures and franchises

No Comments

No Known Activations