INDEX

Explanations

references to the phrase "Hot" with a number attached to it, probably indicating a ranking or rating

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

theless

-0.91

 acknow

-0.75

 literacy

-0.71

 conclud

-0.70

xual

-0.69

ufact

-0.69

 confir

-0.68

 convol

-0.68

 includ

-0.67

 muster

-0.66

POSITIVE LOGITS

Spot

1.04

ness

0.97

fixes

0.96

fix

0.95

shots

0.94

dog

0.92

keys

0.90

mail

0.87

ters

0.87

heads

0.87

Activations Density 0.012%

references to the phrase "Hot" with a number attached to it, probably indicating a ranking or rating

No Comments

No Known Activations