INDEX

Explanations

names of specific individuals and references to popular culture

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

actionDate

-0.86

owship

-0.81

 deepening

-0.80

 contrasting

-0.78

 deleg

-0.77

 conclud

-0.77

 arsen

-0.76

recated

-0.72

PsyNetMessage

-0.71

 tert

-0.70

POSITIVE LOGITS

 Bunny

1.30

 Boys

1.14

 Balls

1.09

 Bear

1.08

 Butt

1.08

 Mouth

1.06

 Guys

1.05

 Pants

1.05

 Productions

1.03

 Girl

1.02

Activations Density 0.238%

names of specific individuals and references to popular culture

No Comments

No Known Activations