INDEX

Explanations

names of people or characters

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

ĸļ

-0.87

enhagen

-0.77

rency

-0.72

choes

-0.71

ruary

-0.70

ocation

-0.65

rences

-0.64

orer

-0.63

fman

-0.62

ngth

-0.62

POSITIVE LOGITS

 Parish

0.80

atari

0.65

 Saints

0.63

 Tale

0.62

alam

0.61

 Lara

0.59

ara

0.57

 Gandhi

0.56

angan

0.56

 "$:/

0.55

Activations Density 0.181%

names of people or characters

No Comments

No Known Activations