INDEX

Explanations

phrases related to deepening or enhancing suspicions or situations

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

inates

-0.69

psc

-0.69

article

-0.65

erity

-0.65

PAR

-0.63

tics

-0.62

alian

-0.62

girl

-0.62

sa

-0.62

gmail

-0.62

POSITIVE LOGITS

 tensions

0.83

 exponentially

0.82

mble

0.77

 hostilities

0.76

 considerably

0.74

 tension

0.72

 markedly

0.72

 dramatically

0.72

 rift

0.69

 downward

0.69

Activations Density 0.200%

phrases related to deepening or enhancing suspicions or situations

No Comments

No Known Activations