INDEX

Explanations

phrases related to media or cultural references

oai_token-act-pair · gpt-3.5-turbo

references to specific conditions or prominent 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

erity

-0.78

 ---------

-0.74

essor

-0.70

hov

-0.69

resents

-0.68

esty

-0.68

llers

-0.68

inge

-0.67

cho

-0.66

ippy

-0.66

POSITIVE LOGITS

ãĥ¥

0.81

 Valhalla

0.77

Saharan

0.70

 Runes

0.64

 registry

0.64

 empires

0.63

 deterrent

0.63

 empire

0.62

 licences

0.62

 spree

0.61

Activations Density 0.000%

phrases related to media or cultural references

references to specific conditions or prominent topics

No Comments

No Known Activations