INDEX

Explanations

references to television networks or stations

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

uate

-0.83

uating

-0.64

forcement

-0.64

arettes

-0.64

vals

-0.64

illard

-0.63

zzle

-0.63

butt

-0.62

estamp

-0.62

ouver

-0.61

POSITIVE LOGITS

 Interstellar

0.85

FIL

0.85

kies

0.83

thood

0.83

kie

0.82

ANY

0.81

NZ

0.80

 Television

0.79

LV

0.79

NI

0.78

Activations Density 0.004%

references to television networks or stations

No Comments

No Known Activations