INDEX

Explanations

mentions of the word "Med" or any of its variations like "Mediator," "Media," or "Medical."

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 root

-0.66

 Nightmare

-0.63

ARD

-0.62

 pity

-0.61

OWER

-0.60

WD

-0.60

 roadside

-0.60

 virginity

-0.60

 Roots

-0.59

 stump

-0.59

POSITIVE LOGITS

ician

1.42

icians

1.36

ical

1.22

icate

1.21

ications

1.20

icals

1.14

ication

1.13

icator

1.13

ically

1.10

icum

1.09

Activations Density 0.895%

mentions of the word "Med" or any of its variations like "Mediator," "Media," or "Medical."

No Comments

No Known Activations