INDEX

Explanations

keywords related to card game mechanics and historical figures

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

lihood

-0.66

ãģį

-0.63

 strawberries

-0.62

REDACTED

-0.60

 Morales

-0.60

 stranger

-0.60

hips

-0.58

-0.56

IVES

-0.56

ive

-0.55

POSITIVE LOGITS

ming

1.41

ilitary

1.37

achine

1.35

pton

1.33

essage

1.32

borgh

1.31

obile

1.28

pson

1.25

icro

1.25

agnetic

1.23

Activations Density 2.220%

keywords related to card game mechanics and historical figures

No Comments

No Known Activations