INDEX

Explanations

phrases related to packaging and products

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

acial

-0.69

inel

-0.67

atio

-0.66

terday

-0.65

psey

-0.63

okin

-0.62

DonaldTrump

-0.62

nesota

-0.62

braska

-0.61

ansson

-0.61

POSITIVE LOGITS

 Contents

1.02

mates

0.92

 containers

0.84

 Crate

0.83

 container

0.82

 containing

0.81

ainers

0.80

 stuffing

0.79

raft

0.78

 overflowing

0.78

Activations Density 2.752%

phrases related to packaging and products

No Comments

No Known Activations