INDEX

Explanations

words related to conveying information effectively

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

anwhile

-0.73

laus

-0.71

night

-0.70

maxwell

-0.68

DAY

-0.61

 Pupp

-0.61

ERAL

-0.58

pmwiki

-0.58

FSA

-0.56

toe

-0.55

POSITIVE LOGITS

ying

1.08

lling

1.04

ller

1.01

structed

1.00

llers

0.96

erers

0.91

ishing

0.90

orld

0.88

xit

0.87

iler

0.86

Activations Density 0.040%

words related to conveying information effectively

No Comments

No Known Activations