INDEX

Explanations

words related to inscriptions or markings

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

arcity

-0.84

HCR

-0.82

APH

-0.80

onym

-0.79

IPP

-0.75

forts

-0.74

onymous

-0.71

IVES

-0.70

izophren

-0.69

IFIED

-0.69

POSITIVE LOGITS

 forehead

1.01

 cheeks

0.98

 windshield

0.92

 cheek

0.91

 abdomen

0.91

 palms

0.90

 sleeves

0.90

 shoulders

0.87

 sleeve

0.85

 thigh

0.84

Activations Density 1.757%

words related to inscriptions or markings

No Comments

No Known Activations