INDEX

Explanations

instances of the word "mad" indicating strong emotional responses

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

ãģ¯

-0.74

sites

-0.69

utical

-0.65

 Islands

-0.65

SPONSORED

-0.64

 distingu

-0.64

 leaflets

-0.62

chnology

-0.61

roma

-0.60

akeru

-0.60

POSITIVE LOGITS

rid

1.03

agascar

1.01

der

1.01

oline

0.90

cap

0.90

rup

0.89

oka

0.86

ras

0.86

cow

0.82

iated

0.82

Activations Density 0.005%

instances of the word "mad" indicating strong emotional responses

No Comments

No Known Activations