INDEX

Explanations

references to television show seasons and episodes

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

hai

-0.74

rity

-0.71

rag

-0.69

mercial

-0.67

tons

-0.65

ogle

-0.65

istani

-0.64

 Haram

-0.64

creen

-0.63

 encl

-0.62

POSITIVE LOGITS

 finale

1.30

 opener

1.27

 premiere

0.95

 unbeaten

0.72

 Playoffs

0.72

 ticket

0.70

 renewal

0.70

é¾

0.69

orsi

0.68

ally

0.68

Activations Density 0.413%

references to television show seasons and episodes

No Comments

No Known Activations