INDEX

Explanations

names or titles related to leadership roles

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ebook

-0.80

ramid

-0.79

onite

-0.78

ograph

-0.77

ech

-0.75

eno

-0.73

antic

-0.72

ographed

-0.72

odi

-0.71

obyl

-0.71

POSITIVE LOGITS

 director

1.04

 editor

1.04

chief

0.99

 chief

0.99

 secretary

0.97

 president

0.96

 Director

0.96

 chairman

0.94

 versa

0.93

director

0.93

Activations Density 0.671%

names or titles related to leadership roles

No Comments

No Known Activations