INDEX

Explanations

phrases related to authority and expertise

oai_token-act-pair · gpt-4o-mini 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 optional

-0.80

 allowance

-0.73

inal

-0.71

ubs

-0.69

 pend

-0.67

urg

-0.67

rogen

-0.67

 pros

-0.66

 bias

-0.66

 warr

-0.66

POSITIVE LOGITS

His

1.28

 Born

1.23

Born

1.17

 Having

1.12

He

1.07

 Located

1.05

 Featuring

1.02

His

1.02

 Firstly

1.01

 Though

0.98

Activations Density 0.497%

phrases related to authority and expertise

No Comments

No Known Activations