INDEX

Explanations

references to certain speculations or hypotheses

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

mouth

-0.77

arak

-0.73

ãĥīãĥ©

-0.73

Rated

-0.71

oses

-0.71

ãģ®éŃĶ

-0.70

ands

-0.70

apsed

-0.70

arez

-0.68

ETS

-0.68

POSITIVE LOGITS

 maybe

0.95

 perhaps

0.95

 someday

0.80

 whoever

0.80

perhaps

0.75

might

0.74

yip

0.72

 although

0.72

 disgruntled

0.70

 motive

0.67

Activations Density 0.155%

references to certain speculations or hypotheses

No Comments

No Known Activations