INDEX

Explanations

phrases related to historical wartime figures or events

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

schild

-0.83

STEP

-0.81

lihood

-0.74

heartedly

-0.70

thens

-0.70

sonian

-0.69

merce

-0.68

ycle

-0.67

bom

-0.65

theless

-0.65

POSITIVE LOGITS

mone

1.08

utations

1.02

utation

0.99

azar

0.92

zman

0.87

omon

0.87

uting

0.86

isbury

0.85

utes

0.84

erno

0.83

Activations Density 0.831%

phrases related to historical wartime figures or events

No Comments

No Known Activations