INDEX

Explanations

words related to negative events or misdeeds

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

 Fres

-0.69

flix

-0.66

raltar

-0.64

belt

-0.63

retty

-0.62

 cooled

-0.62

 Ferr

-0.61

 Irma

-0.60

hower

-0.60

 Chest

-0.59

POSITIVE LOGITS

vous

1.00

ceptions

0.90

gotten

0.88

ceived

0.84

akable

0.82

dule

0.80

imate

0.79

ventures

0.78

rued

0.78

ception

0.76

Activations Density 0.023%

words related to negative events or misdeeds

No Comments

No Known Activations