INDEX

Explanations

phrases related to various locations or elements in a setting

oai_token-act-pair · gpt-3.5-turbo

locations and settings within narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

amia

-0.79

 Canaver

-0.78

urations

-0.73

olitics

-0.71

conserv

-0.70

cies

-0.69

ï¸ı

-0.68

hops

-0.68

TRUMP

-0.68

baugh

-0.67

POSITIVE LOGITS

 skyline

1.05

 floor

0.98

 clipboard

0.96

 kitchen

0.92

 horizon

0.91

 machinery

0.91

 crest

0.89

 atmosphere

0.87

 cafeteria

0.86

 logo

0.86

Activations Density 0.580%

phrases related to various locations or elements in a setting

locations and settings within narratives

No Comments

No Known Activations