INDEX

Explanations

phrases related to specific locations or names, particularly with the word "jar"

oai_token-act-pair · gpt-3.5-turbo

references to jars or similar containers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ingham

-0.79

BOOK

-0.68

 exerc

-0.68

 Eleven

-0.66

ãģį

-0.66

states

-0.65

ATK

-0.65

 Eater

-0.65

 defe

-0.65

arettes

-0.59

POSITIVE LOGITS

jar

0.96

iasis

0.83

zan

0.81

anism

0.80

bled

0.79

akia

0.79

nesses

0.79

azz

0.78

anian

0.78

oslov

0.77

Activations Density 0.017%

phrases related to specific locations or names, particularly with the word "jar"

references to jars or similar containers

No Comments

No Known Activations