INDEX

Explanations

phrases related to programming languages and code behavior

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

inav

-0.78

 senses

-0.69

 wills

-0.65

 membranes

-0.63

 traged

-0.61

eor

-0.61

peg

-0.60

 Notting

-0.60

 Awakens

-0.59

 elders

-0.58

POSITIVE LOGITS

rael

1.21

senal

1.06

omorphic

1.02

nt

1.01

 hereby

1.00

olation

0.99

 gonna

0.98

ometric

0.95

olate

0.92

abella

0.89

Activations Density 0.647%

phrases related to programming languages and code behavior

No Comments

No Known Activations