INDEX

Explanations

references to negative or catastrophic events or situations

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

agine

-0.83

trl

-0.81

sure

-0.77

yi

-0.77

dinand

-0.76

uador

-0.74

andise

-0.74

uana

-0.74

ynthesis

-0.74

theless

-0.73

POSITIVE LOGITS

 offender

1.08

 offenders

1.08

 imaginable

0.97

 behaved

0.94

 nightmare

0.84

 possible

0.79

 performer

0.78

 nightmares

0.76

 conceivable

0.74

 hitter

0.73

Activations Density 0.023%

references to negative or catastrophic events or situations

No Comments

No Known Activations