INDEX

Explanations

informational items or specific keywords related to different topics

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

 Stones

-0.84

 Thrones

-0.75

 Canaver

-0.74

 Party

-0.74

 Legions

-0.73

 Yard

-0.73

 Inspection

-0.72

 Corps

-0.71

 Bundy

-0.70

 Reconstruction

-0.69

POSITIVE LOGITS

etheless

1.02

terday

1.01

theless

0.97

ande

0.93

0.91

0.90

anwhile

0.88

bsite

0.86

yond

0.84

fortunately

0.83

Activations Density 1.453%

informational items or specific keywords related to different topics

No Comments

No Known Activations