INDEX

Explanations

names of individuals

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

milo

-0.78

adobe

-0.65

 suspic

-0.64

Brazil

-0.64

Downloadha

-0.63

pmwiki

-0.62

ESSION

-0.59

âĸ¬

-0.56

 TEST

-0.55

 credential

-0.55

POSITIVE LOGITS

 alike

1.30

 respectively

1.03

eele

0.85

axter

0.83

LLP

0.82

 striped

0.71

 Streets

0.69

 Productions

0.66

 Circus

0.66

 together

0.65

Activations Density 0.181%

names of individuals

No Comments

No Known Activations