INDEX

Explanations

information about events and incidents, such as accidents, political statements, and TV show renewals

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

}.

-0.57

*.

-0.57

 thereto

-0.55

 respectively

-0.55

 accordingly

-0.54

 thereof

-0.53

+.

-0.53

>.

-0.53

 versa

-0.53

_.

-0.52

POSITIVE LOGITS

 Canaver

0.52

 crochet

0.42

 Guant

0.42

 OnePlus

0.39

 Bitcoin

0.38

 Orioles

0.38

NBN

0.37

 Grateful

0.36

 cryptocurrency

0.35

Â¶

0.35

Activations Density 23.313%

information about events and incidents, such as accidents, political statements, and TV show renewals

No Comments

No Known Activations