INDEX

Explanations

titles and roles within organizations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

à©

-0.61

âĵĺ

-0.61

 therapists

-0.61

 Turks

-0.60

 slopes

-0.58

 unlaw

-0.58

 brackets

-0.57

uld

-0.56

 clips

-0.55

 strangers

-0.55

POSITIVE LOGITS

 overseeing

0.91

itol

0.70

0.69

 extraord

0.69

 emer

0.67

 coordinating

0.65

 credited

0.65

stration

0.64

zona

0.64

ews

0.64

Activations Density 0.107%

titles and roles within organizations

No Comments

No Known Activations