INDEX

Explanations

names of individuals and organizations

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

ufact

-0.71

vertisement

-0.63

odder

-0.62

ICAN

-0.62

erity

-0.61

FG

-0.61

yright

-0.60

cot

-0.60

leneck

-0.58

Ã£

-0.58

POSITIVE LOGITS

utsche

0.74

pta

0.71

pret

0.66

onica

0.66

aternity

0.64

 Dull

0.64

deen

0.63

lde

0.63

oslav

0.60

Ae

0.59

Activations Density 0.413%

names of individuals and organizations

No Comments

No Known Activations