INDEX

Explanations

mentions of the word "Holy" followed by another word

oai_token-act-pair · gpt-3.5-turbo

references to the term "Holy" in various contexts

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

PsyNetMessage

-0.88

phrine

-0.87

ickr

-0.84

atown

-0.79

yrinth

-0.76

unct

-0.73

ĸļ

-0.73

apses

-0.72

urable

-0.72

POSITIVE LOGITS

 Grail

1.00

 Scripture

0.74

 Trinity

0.73

field

0.72

 Maiden

0.70

gate

0.70

father

0.69

 communion

0.66

 Orbit

0.66

 Gorge

0.66

Activations Density 0.011%

mentions of the word "Holy" followed by another word

references to the term "Holy" in various contexts

No Comments

No Known Activations