INDEX

Explanations

references to weight measurements

oai_token-act-pair · gpt-4o-mini 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

opol

-0.86

choes

-0.76

istically

-0.76

 externalToEVAOnly

-0.72

effects

-0.71

cknow

-0.70

eer

-0.69

vironment

-0.68

 Nare

-0.66

icent

-0.66

POSITIVE LOGITS

 sterling

1.23

 heavier

0.95

lbs

0.94

 gorilla

0.94

 overweight

0.92

 pounds

0.90

cake

0.88

lbs

0.88

 weight

0.88

Weight

0.84

Activations Density 0.016%

references to weight measurements

No Comments

No Known Activations