INDEX

Explanations

terms related to the National Geographic magazine

oai_token-act-pair · gpt-3.5-turbo

references to National Geographic and associated entities or topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

unts

-0.86

rahim

-0.85

bub

-0.82

ework

-0.79

ammy

-0.75

olulu

-0.75

rb

-0.74

raising

-0.73

nea

-0.72

ihara

-0.71

POSITIVE LOGITS

 Geographic

1.00

 Warfare

0.83

 Channel

0.83

 Explorer

0.77

 Editors

0.72

 explorer

0.72

 inference

0.71

 VIDEOS

0.71

 Dispatch

0.70

 Correspond

0.70

Activations Density 0.009%

terms related to the National Geographic magazine

references to National Geographic and associated entities or topics

No Comments

No Known Activations