INDEX

Explanations

locations, specifically focusing on Manila

oai_token-act-pair · gpt-3.5-turbo

references to Manila and its surrounding context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Ö¼

-0.85

chool

-0.84

Ã¸

-0.81

ãĥ¤

-0.79

ordan

-0.79

RH

-0.76

ãĥł

-0.73

heimer

-0.72

 Macintosh

-0.72

URN

-0.70

POSITIVE LOGITS

 Manila

1.01

 Filip

0.95

 Filipino

0.89

 Marcos

0.87

 Duterte

0.87

pes

0.86

 Philippines

0.85

 typh

0.84

atown

0.83

 Philippine

0.82

Activations Density 0.012%

locations, specifically focusing on Manila

references to Manila and its surrounding context

No Comments

No Known Activations