INDEX

Explanations

phrases related to degrees or classifications of intensity or seriousness

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orial

-0.78

enegger

-0.70

 Andromeda

-0.69

ansson

-0.67

 Leilan

-0.67

luaj

-0.67

oris

-0.65

oras

-0.65

LM

-0.65

ãĥĹ

-0.64

POSITIVE LOGITS

 assurance

0.91

xual

0.78

 deviation

0.77

payers

0.75

 shif

0.74

 crossings

0.74

 levels

0.73

 difference

0.72

etting

0.70

 advancement

0.67

Activations Density 2.055%

phrases related to degrees or classifications of intensity or seriousness

No Comments

No Known Activations