INDEX

Explanations

phrases related to debunking myths or falsehoods

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

meal

-0.86

faces

-0.72

hetical

-0.66

istics

-0.65

hens

-0.65

face

-0.65

 Inquisitor

-0.63

velt

-0.61

milo

-0.59

 shotguns

-0.59

POSITIVE LOGITS

ilitation

1.14

unker

1.11

ilit

1.08

rief

1.07

orah

1.07

utable

1.06

ounced

1.02

ilitating

0.97

ouncing

0.97

ugg

0.95

Activations Density 0.020%

phrases related to debunking myths or falsehoods

No Comments

No Known Activations