INDEX

Explanations

names of individuals or specific entities

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

ttes

-0.68

 inference

-0.68

 acids

-0.67

 strawberries

-0.67

 strikeouts

-0.65

lihood

-0.65

hani

-0.65

 wool

-0.65

REDACTED

-0.63

 Morales

-0.61

POSITIVE LOGITS

ming

1.52

essage

1.39

borgh

1.35

pering

1.33

pered

1.30

ilitary

1.27

mer

1.27

pton

1.24

ajor

1.23

mers

1.22

Activations Density 4.654%

names of individuals or specific entities

No Comments

No Known Activations