INDEX

Explanations

proper nouns, particularly titles and names

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

creen

-0.71

grass

-0.70

 Schumer

-0.65

 Libre

-0.62

SOS

-0.62

 slack

-0.61

 Mandela

-0.60

low

-0.60

 Sons

-0.60

terson

-0.60

POSITIVE LOGITS

haps

1.38

ansom

1.37

outine

1.31

ascal

1.31

senal

1.30

itchie

1.23

ifles

1.19

ework

1.16

otation

1.15

otten

1.14

Activations Density 0.329%

proper nouns, particularly titles and names

No Comments

No Known Activations