INDEX

Explanations

mentions of the specific word "Silver" with different variations in the formatting

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

etically

-1.01

uador

-0.79

etic

-0.74

etics

-0.72

 Advertisement

-0.70

ynt

-0.70

 Gutenberg

-0.68

NEY

-0.67

lov

-0.66

urable

-0.66

POSITIVE LOGITS

iod

1.08

moon

0.98

stein

0.98

anguage

0.97

blind

0.94

mont

0.92

ware

0.89

fish

0.88

stone

0.87

beard

0.86

Activations Density 0.013%

mentions of the specific word "Silver" with different variations in the formatting

No Comments

No Known Activations