INDEX

Explanations

mentions of specific terms or acronyms

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lain

-0.71

ienced

-0.67

ividual

-0.67

rity

-0.66

manship

-0.64

 narrator

-0.62

 Norn

-0.62

 glim

-0.61

itated

-0.60

 innocence

-0.59

POSITIVE LOGITS

zzi

1.51

ogle

1.29

zzo

1.26

ppo

1.24

opy

1.23

pper

1.22

zzle

1.20

ilers

1.18

pping

1.16

oby

1.16

Activations Density 1.966%

mentions of specific terms or acronyms

No Comments

No Known Activations