INDEX

Explanations

references to specific logos or emblems

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

kus

-0.78

hower

-0.74

neau

-0.72

theless

-0.72

hurst

-0.72

cible

-0.65

Ago

-0.64

 Lank

-0.63

dose

-0.61

ikan

-0.61

POSITIVE LOGITS

 logos

1.02

 logo

0.93

 Doodle

0.89

 Logo

0.87

 slogan

0.82

plates

0.82

 recognition

0.79

 trademarks

0.77

tag

0.77

 synonymous

0.75

Activations Density 0.053%

references to specific logos or emblems

No Comments

No Known Activations