INDEX

Explanations

references to recipients or receivers of various actions or objects

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

ERO

-0.75

ucci

-0.70

 slic

-0.67

udic

-0.67

 Nanto

-0.65

 Sabb

-0.64

Ku

-0.63

anova

-0.60

kill

-0.59

 Zimmerman

-0.58

POSITIVE LOGITS

 recipients

0.86

ients

0.85

 recipient

0.84

ipient

0.82

soever

0.77

ted

0.75

ience

0.74

imental

0.71

rocal

0.71

cffff

0.70

Activations Density 0.033%

references to recipients or receivers of various actions or objects

No Comments

No Known Activations