INDEX

Explanations

mentions of the term "Full"

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

affles

-0.70

mone

-0.69

abbit

-0.66

 Tycoon

-0.66

akov

-0.65

pher

-0.65

 Disciple

-0.64

 whis

-0.64

ovan

-0.62

Downloadha

-0.61

POSITIVE LOGITS

erton

1.12

screen

0.98

 blown

0.96

 fled

0.96

 complement

0.94

fledged

0.90

blown

0.89

 frontal

0.88

 disclosure

0.84

ness

0.84

Activations Density 1.093%

mentions of the term "Full"

No Comments

No Known Activations