INDEX

Explanations

the number "two"

oai_token-act-pair · gpt-3.5-turbo

instances or references to the number two 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

enment

-0.73

assisted

-0.67

ãĥ¼ãĥ«

-0.67

Attempts

-0.66

buster

-0.64

evaluate

-0.63

venture

-0.63

alloc

-0.63

lean

-0.62

)=(

-0.61

POSITIVE LOGITS

 reason

0.80

 shortage

0.78

 reasons

0.77

HUN

0.72

 chance

0.71

WAY

0.71

 apiece

0.67

 vacancies

0.67

 distinct

0.66

 similarities

0.66

Activations Density 0.075%

the number "two"

instances or references to the number two in various contexts

No Comments

No Known Activations