INDEX

Explanations

specific comparisons for evaluation

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

redo

-0.69

perm

-0.68

ERY

-0.64

AIR

-0.64

ISTER

-0.63

tank

-0.59

OIL

-0.58

KER

-0.57

guard

-0.56

~~~~

-0.56

POSITIVE LOGITS

 dots

1.15

 favorably

1.02

 disparate

0.85

MpServer

0.85

 existing

0.75

 negatively

0.72

 geographically

0.70

 perspectives

0.67

 insult

0.66

 sexes

0.66

Activations Density 0.174%

specific comparisons for evaluation

No Comments

No Known Activations