INDEX

Explanations

points or items in a list that emphasize and support arguments

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

erity

-0.77

roth

-0.73

cffffcc

-0.71

Ń·

-0.71

anship

-0.69

ipment

-0.69

izont

-0.68

status

-0.68

leases

-0.68

own

-0.67

POSITIVE LOGITS

 Highly

0.66

 Reasons

0.66

 Helpful

0.64

âĺħ

0.62

 vegetarian

0.62

 bestselling

0.61

 debunk

0.60

ottest

0.59

å°Ĩ

0.58

 recommend

0.58

Activations Density 0.083%

points or items in a list that emphasize and support arguments

No Comments

No Known Activations