INDEX

Explanations

phrases related to specific nouns or names, often starting with a capital letter

oai_token-act-pair · gpt-3.5-turbo

references to specific names and terms associated with notable movies or characters

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¬

-0.87

 indo

-0.82

hetic

-0.66

iments

-0.66

 partName

-0.65

FTWARE

-0.65

âĸ¬âĸ¬

-0.65

hetically

-0.64

iment

-0.63

eco

-0.62

POSITIVE LOGITS

shots

0.81

iewicz

0.80

Leaks

0.73

antha

0.71

roots

0.71

lord

0.71

shot

0.65

urrection

0.64

 Wick

0.63

steps

0.63

Activations Density 0.052%

phrases related to specific nouns or names, often starting with a capital letter

references to specific names and terms associated with notable movies or characters

No Comments

No Known Activations