INDEX

Explanations

instances of '&' with surrounding letters or numbers

oai_token-act-pair · gpt-3.5-turbo

occurrences of the ampersand character (&)

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

 Nebula

-0.87

etheless

-0.80

owicz

-0.77

orious

-0.70

ogene

-0.69

eus

-0.67

 Luther

-0.66

izen

-0.65

agher

-0.64

 Mara

-0.63

POSITIVE LOGITS

Bs

0.77

Cs

0.77

RT

0.77

RD

0.76

Co

0.76

HI

0.75

FAQ

0.75

idd

0.74

cas

0.73

IQ

0.72

Activations Density 0.011%

instances of '&' with surrounding letters or numbers

occurrences of the ampersand character (&)

No Comments

No Known Activations