INDEX

Explanations

mentions of a specific location - Ottawa

oai_token-act-pair · gpt-3.5-turbo

mentions of the city Ottawa

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

JD

-0.67

 stere

-0.64

 jars

-0.62

 missiles

-0.61

 eclipse

-0.60

mir

-0.60

 defect

-0.60

 missile

-0.59

 conglomer

-0.59

dex

-0.59

POSITIVE LOGITS

 Ottawa

3.84

Ott

3.14

 Edmonton

2.34

 Canberra

2.28

 Winnipeg

2.27

 Calgary

2.23

 Montreal

2.09

Toronto

2.08

 Saskatchewan

2.07

 Toronto

2.05

Activations Density 0.015%

mentions of a specific location - Ottawa

mentions of the city Ottawa

No Comments

No Known Activations