INDEX

Explanations

terms related to television news and media services

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

tein

-0.96

heit

-0.87

oleon

-0.73

redo

-0.67

onomous

-0.67

ierre

-0.67

yang

-0.64

maxwell

-0.64

 Flask

-0.63

ullah

-0.63

POSITIVE LOGITS

levision

0.91

 playback

0.88

 telev

0.80

 broadcasting

0.79

DVD

0.78

 commercials

0.77

 studios

0.76

elight

0.76

 movies

0.74

tainment

0.73

Activations Density 0.176%

terms related to television news and media services

No Comments

No Known Activations