INDEX

Explanations

superlatives describing the highest or most significant entities in diverse categories

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ADRA

-0.87

ĪĴ

-0.83

Module

-0.73

odge

-0.73

zes

-0.70

regon

-0.70

escription

-0.67

ints

-0.67

okane

-0.67

azon

-0.67

POSITIVE LOGITS

 EVER

0.71

behind

0.70

âĦ¢:

0.70

 according

0.69

 standings

0.68

 today

0.67

âĦ¢

0.66

(~

0.66

 besides

0.65

0.64

Activations Density 0.123%

superlatives describing the highest or most significant entities in diverse categories

No Comments

No Known Activations