INDEX

Explanations

technical instructions or processes

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

initions

-0.70

 Lands

-0.67

 designation

-0.66

HOME

-0.66

NCT

-0.63

gdala

-0.63

 moniker

-0.62

 Waves

-0.62

 Dragons

-0.62

 Generator

-0.61

POSITIVE LOGITS

ably

1.23

ificantly

1.20

ibly

1.20

istically

1.19

ivable

1.18

ishly

1.18

ifully

1.17

iably

1.16

etically

1.15

vable

1.14

Activations Density 1.682%

technical instructions or processes

No Comments

No Known Activations