INDEX

Explanations

references to news and photography

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ressive

-0.82

ensable

-0.76

asons

-0.74

astically

-0.72

plot

-0.71

yrinth

-0.69

Mos

-0.67

ricular

-0.67

schild

-0.67

»Ĵ

-0.66

POSITIVE LOGITS

 caption

0.81

 IMAGES

0.79

 Photographer

0.70

FILE

0.69

 Phot

0.67

 circulating

0.67

 REUTERS

0.66

 Nare

0.66

 photo

0.65

 circa

0.64

Activations Density 0.035%

references to news and photography

No Comments

No Known Activations