INDEX

Explanations

words related to historical events, government actions, and political concepts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

specific phrases and terms related to actions and events in narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

 TAMADRA

-0.72

nih

-0.64

adish

-0.64

inent

-0.63

urated

-0.63

inite

-0.63

oded

-0.63

etermination

-0.60

ode

-0.59

ICH

-0.59

POSITIVE LOGITS

 back

2.47

back

2.41

 Back

2.27

 backs

2.19

backs

2.17

Back

2.14

 BACK

2.08

BACK

2.02

 backend

1.28

 backward

1.15

Activations Density 0.448%

words related to historical events, government actions, and political concepts

specific phrases and terms related to actions and events in narratives

No Comments

No Known Activations