INDEX

Explanations

instances referencing the concept of finality or completion

oai_token-act-pair · gpt-3.5-turbo

instances of the word "final" and its related 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

 Spears

-0.68

men

-0.67

LI

-0.65

Kuh

-0.64

 charism

-0.64

holes

-0.62

ris

-0.61

stadt

-0.60

spr

-0.60

 ministries

-0.59

POSITIVE LOGITS

 FANTASY

0.87

iminary

0.82

ife

0.82

izers

0.78

ITY

0.78

ebin

0.78

izing

0.78

aneously

0.76

istically

0.76

 showdown

0.76

Activations Density 0.021%

instances referencing the concept of finality or completion

instances of the word "final" and its related contexts

No Comments

No Known Activations