INDEX

Explanations

calls to action related to watching videos or shows

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

bun

-0.70

 amalg

-0.57

 turf

-0.57

nib

-0.56

 assum

-0.56

 recl

-0.55

 barley

-0.55

 bisc

-0.55

 peanuts

-0.54

 clay

-0.54

POSITIVE LOGITS

 WATCH

0.87

 VIDEOS

0.73

 Watching

0.70

 Videos

0.68

esome

0.67

dog

0.65

 View

0.64

 Sasha

0.63

...]

0.61

 Canaveral

0.61

Activations Density 0.001%

calls to action related to watching videos or shows

No Comments

No Known Activations