INDEX

Explanations

references to award nominations and winners in various categories, particularly at the Oscars

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

Negative Logits

Uz

-0.81

CDC

-0.77

ALTH

-0.75

 Anon

-0.74

gun

-0.72

itness

-0.69

DEP

-0.64

rir

-0.64

Beg

-0.64

meal

-0.62

POSITIVE LOGITS

 nominations

1.37

 nominees

1.21

 nomination

1.11

 Awards

1.10

 nominating

1.02

 awards

1.02

 Oscars

1.01

 nominated

1.01

 finalists

0.98

 winner

0.97

Activations Density 0.047%

references to award nominations and winners in various categories, particularly at the Oscars

No Known Activations