INDEX

Explanations

references to popularity or public approval

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

thur

-0.71

 Centauri

-0.68

RAW

-0.68

ERO

-0.68

©¶æ

-0.67

ural

-0.67

agher

-0.66

 Aviv

-0.66

ignt

-0.64

ĸļ

-0.62

POSITIVE LOGITS

ized

1.14

izing

1.09

ity

1.07

ised

1.03

izations

1.00

izers

0.96

ization

0.95

izer

0.92

isations

0.92

ize

0.90

Activations Density 0.029%

references to popularity or public approval

No Comments

No Known Activations