INDEX

Explanations

phrases related to the term "Fu" at different activation levels

oai_token-act-pair · gpt-3.5-turbo

references to the name or term "Fu."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Cosponsors

-0.85

 Remastered

-0.77

 Divide

-0.70

Ö¼

-0.69

 Volunte

-0.68

 Emin

-0.68

 Corinth

-0.66

 Metatron

-0.66

 Barrett

-0.65

*/(

-0.63

POSITIVE LOGITS

cci

1.42

selage

1.35

lda

1.22

pport

1.21

pload

1.16

isine

1.15

cca

1.11

pper

1.08

lder

1.07

ppy

1.04

Activations Density 0.045%

phrases related to the term "Fu" at different activation levels

references to the name or term "Fu."

No Comments

No Known Activations