INDEX

Explanations

phrases related to geographic locations or physical features

oai_token-act-pair · gpt-3.5-turbo

terms related to geographic locations and infrastructure

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

 personalized

-0.70

aughs

-0.67

 Pets

-0.66

 Answers

-0.64

 Apocalypse

-0.64

 Ladies

-0.64

ombies

-0.62

 Zombies

-0.62

VERSION

-0.62

 Summers

-0.62

POSITIVE LOGITS

 corridor

1.33

 boundary

1.28

 junction

1.25

 zone

1.25

 separating

1.18

 fence

1.17

 barrier

1.12

 canyon

1.12

 perimeter

1.09

 intersection

1.09

Activations Density 0.312%

phrases related to geographic locations or physical features

terms related to geographic locations and infrastructure

No Comments

No Known Activations