INDEX

Explanations

phrases contrasting opposing views or arguments

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iard

-0.83

ranged

-0.75

tesy

-0.73

mail

-0.72

dropping

-0.70

saw

-0.70

noon

-0.68

ahead

-0.68

rough

-0.68

MpServer

-0.67

POSITIVE LOGITS

 beliefs

1.01

 reality

0.99

 expectations

0.95

 belief

0.94

 tenets

0.91

 orthodoxy

0.89

 prevailing

0.88

 worldview

0.88

 dogma

0.87

 ideals

0.87

Activations Density 0.053%

phrases contrasting opposing views or arguments

No Comments

No Known Activations