INDEX

Explanations

questions and answers

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Leban

-0.69

wana

-0.69

erker

-0.66

robat

-0.64

awar

-0.64

 Vengeance

-0.64

 Dull

-0.63

 Nanto

-0.63

DAQ

-0.63

akin

-0.63

POSITIVE LOGITS

answer

0.96

 answ

0.95

ysis

0.94

 answer

0.91

 thereto

0.90

answered

0.89

swers

0.87

 answered

0.84

 answers

0.80

naires

0.78

Activations Density 0.022%

questions and answers

No Comments

No Known Activations