INDEX

Explanations

statements related to electronic devices and advertising

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

 TAMADRA

-0.74

 Templar

-0.74

hani

-0.70

claw

-0.69

quartered

-0.69

lihood

-0.68

iciary

-0.66

 Apostles

-0.66

uble

-0.65

mates

-0.65

POSITIVE LOGITS

 advertising

1.01

 banners

0.96

 campaigns

0.96

vertising

0.95

verts

0.93

 billboards

0.92

 touting

0.92

strip

0.88

 blitz

0.86

 Advertising

0.85

Activations Density 1.154%

statements related to electronic devices and advertising

No Comments

No Known Activations