INDEX

Explanations

locations, specifically different places and cities

oai_token-act-pair · gpt-3.5-turbo

geographical locations and specific places

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

ogens

-0.75

measures

-0.73

winner

-0.73

verbs

-0.71

ocious

-0.71

apter

-0.71

bons

-0.68

ordan

-0.67

 sacrific

-0.65

DERR

-0.64

POSITIVE LOGITS

 apartment

1.18

 courthouse

1.15

 mansion

1.13

 office

1.11

 hotel

1.10

 facility

1.09

 airport

1.08

 warehouse

1.08

 offices

1.06

 suburb

1.06

Activations Density 0.126%

locations, specifically different places and cities

geographical locations and specific places

No Comments

No Known Activations