INDEX

Explanations

phrases related to conflicts or disputes

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

sbm

-0.82

tera

-0.78

okemon

-0.75

PsyNetMessage

-0.72

UME

-0.71

 Cosponsors

-0.70

qt

-0.70

 Deaths

-0.69

-0.67

itcher

-0.67

POSITIVE LOGITS

 nutritious

1.10

 prosperous

1.01

 orderly

1.01

 tasty

1.01

 stable

0.97

 efficient

0.97

 tidy

0.96

 healthy

0.96

 unbeat

0.95

 spacious

0.94

Activations Density 0.139%

phrases related to conflicts or disputes

No Comments

No Known Activations