INDEX

Explanations

words related to discussions or arguments, possibly of a confrontational nature

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

 misunder

-0.76

atche

-0.75

senal

-0.69

 mattered

-0.67

 phased

-0.62

etsk

-0.62

iguous

-0.60

 identifiable

-0.60

 concentrating

-0.58

 outlawed

-0.58

POSITIVE LOGITS

 Friends

0.83

nect

0.82

GAN

0.80

Co

0.79

optional

0.79

^^^^

0.74

wcsstore

0.71

 Company

0.68

ÃĽ

0.67

 Associates

0.67

Activations Density 0.031%

words related to discussions or arguments, possibly of a confrontational nature

No Comments

No Known Activations