INDEX

Explanations

numeric values in a structured format

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

ãĥ¤

-0.89

rous

-0.76

irrel

-0.73

acements

-0.73

rade

-0.72

rians

-0.71

itionally

-0.69

anson

-0.68

endez

-0.68

ition

-0.67

POSITIVE LOGITS

teenth

1.30

teen

1.20

patrick

0.75

ppy

0.71

 Aman

0.68

------------------------------------------------

0.67

0.66

ppel

0.66

mos

0.65

Activations Density 1.150%

numeric values in a structured format

No Comments

No Known Activations