INDEX

Explanations

phrases related to the type or category of something

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 VIDEOS

-0.76

edia

-0.69

Kod

-0.62

 Bots

-0.61

oulos

-0.60

UNCH

-0.58

 Rings

-0.58

WOR

-0.58

agnar

-0.57

 Bars

-0.57

POSITIVE LOGITS

etting

1.02

face

0.99

hearted

0.87

casting

0.86

lier

0.83

faces

0.83

lihood

0.80

etter

0.79

ahead

0.77

gom

0.76

Activations Density 1.611%

phrases related to the type or category of something

No Comments

No Known Activations