INDEX

Explanations

statements of strength or strong emphasis

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

 Incarnation

-0.72

 Kare

-0.68

Hop

-0.67

 Wonderland

-0.66

 Marathon

-0.66

 Correction

-0.64

 Newly

-0.63

adr

-0.63

oleon

-0.62

 Hilton

-0.62

POSITIVE LOGITS

nesses

0.94

ener

0.92

 enough

0.88

ament

0.88

 cryptography

0.81

 winds

0.77

 emphasis

0.75

 motiv

0.74

 deterrent

0.74

 coupling

0.73

Activations Density 1.900%

statements of strength or strong emphasis

No Comments

No Known Activations