INDEX

Explanations

occurrences of the word "out."

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

Login

-0.70

olesterol

-0.65

»Ĵ

-0.64

ille

-0.63

hedral

-0.62

anooga

-0.62

TON

-0.62

gee

-0.62

 rede

-0.61

shaw

-0.60

POSITIVE LOGITS

fitted

1.05

wards

0.84

smart

0.83

doors

0.80

skirts

0.78

door

0.70

fitting

0.69

 valves

0.69

bath

0.69

 BALL

0.69

Activations Density 0.014%

occurrences of the word "out."

No Comments

No Known Activations