INDEX

Explanations

phrases related to examining or analyzing something closely

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

apo

-0.74

âĵĺ

-0.67

rio

-0.67

 âĵĺ

-0.66

Cry

-0.65

Lew

-0.65

Skill

-0.64

KO

-0.64

Guest

-0.64

Chel

-0.63

POSITIVE LOGITS

 carefully

0.97

 history

0.91

 objectively

0.91

 histories

0.89

 positives

0.86

 scriptures

0.83

 retrospect

0.83

history

0.83

 literature

0.83

 closely

0.82

Activations Density 1.698%

phrases related to examining or analyzing something closely

No Comments

No Known Activations