INDEX

Explanations

government officials and political terms

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

nikov

-0.89

MpServer

-0.80

crew

-0.79

Äĩ

-0.79

ensis

-0.74

eros

-0.73

zzle

-0.71

mere

-0.71

EStreamFrame

-0.70

eat

-0.68

POSITIVE LOGITS

 Barack

0.93

 George

0.92

 Bill

0.92

 Stephen

0.92

Tim

0.88

Ann

0.88

 Donald

0.88

 David

0.87

 Steve

0.87

 Mitch

0.85

Activations Density 0.117%

government officials and political terms

No Comments

No Known Activations