INDEX

Explanations

proper nouns related to people or places

oai_token-act-pair · gpt-3.5-turbo

references to specific wards or districts within a governing structure

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.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

ãĥĪ

-0.89

ãĥ¬

-0.82

ãĥ©

-0.81

ãĥīãĥ©

-0.78

ctive

-0.73

kered

-0.72

ãĥ³ãĤ¸

-0.70

izations

-0.70

ivia

-0.70

ctions

-0.69

POSITIVE LOGITS

robe

2.15

ens

1.22

wick

1.02

ell

1.02

ynski

1.01

een

0.99

minster

0.95

 Churchill

0.94

lock

0.92

er

0.90

Activations Density 0.046%

proper nouns related to people or places

references to specific wards or districts within a governing structure

No Comments

No Known Activations