INDEX

Explanations

opinions and thoughts expressed by individuals in a discussion or interview setting

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

perty

-0.67

API

-0.66

irement

-0.64

 Yourself

-0.64

thinkable

-0.62

Submit

-0.62

worldly

-0.62

eatured

-0.61

ilaterally

-0.61

foreseen

-0.59

POSITIVE LOGITS

 whoever

0.74

it

0.74

 anecd

0.71

 thats

0.71

 anybody

0.68

 there

0.67

 everybody

0.67

 saddened

0.66

 Cantor

0.66

Activations Density 1.182%

opinions and thoughts expressed by individuals in a discussion or interview setting

No Comments

No Known Activations